Search

The lesion area segmentation technology in endoscopic images helps doctors discover and locate lesion areas, which is crucial for improving the survival rate of patients. In various tasks within the field of medical image segmentation, the fully convolutional network with U-shaped architecture has achieved great success, but the fully convolutional operation has limitations in obtaining global context information.The Segment Anything Model (SAM), the latest foundational model in computer vision, boasts powerful image feature extraction capabilities and can capture explicit long-range dependencies, but it cannot capture low-level detailed information. This paper proposes a dual-branch model, named FCSAM, which makes up for the respective shortcomings of the fully convolutional network and SAM in medical image segmentation. At the same time, this paper proposes a LayerNorm and low-rank-based(LNLoRA) fine-tuning strategy, which overcomes the problem that large models cannot be applied to small-scale medical image datasets and reduces the deployment and storage overhead of the model. In addition, we also design a boundary aware module and a boundary key point map generation algorithm to improve the boundary localization ability of this model, and significantly improve the segmentation performance of lesions. The experimental results show that the IOU, Dice coefficient and accuracy of the model are 79.3%, 88.1% and 93.2% respectively. These metrics show that the model proposed in this paper achieves superior performance to various competing methods in terms of segmentation accuracy and has great potential in clinical analysis and diagnosis.

With the widespread availability and utilization of 3D meshes across various applications, the need for accurately assessing their visual quality has become increasingly important. Despite the significance of this task, the literature offers few No-Reference (NR) approaches for evaluating the visual quality of 3D meshes. In response to this gap, this paper proposes a novel NR approach tailored specifically to score the quality of 3D meshes. After rendering a 3D mesh into 2D views and patches, a pre-trained convolutional neural network automatically extracts deep features from. These features are then employed in a Multi-Layer Perceptron regressor to predict the quality score of the rendered images. The obtained scores are combined with their corresponding BRISQUE scores, and an additional MLP regressor is used to predict the final score. We present experimental results demonstrating the effectiveness and robustness of our approach across a diverse range of 3D mesh datasets. Comparative analyses with existing NR methods underscore the superior performance and versatility of the proposed approach. Overall, this paper contributes to the advancement of NR techniques for assessing 3D mesh quality, offering a valuable tool for researchers, practitioners, and developers working with 3D models across various domains.

As the economy and society continue to develop, the types of underwater targets are becoming increasingly diverse, accompanied by a corresponding increase in environmental noise. This noise presents a significant challenge for the recognition of underwater target radiation noise, as it can result in the loss of effective information during the noise processing stage. This, in turn, reduces the accuracy of underwater target radiation noise recognition. This paper proposes an underwater target radiation noise recognition method based on the enhanced image method. The method transforms the original underwater target radiation noise signal into an enhanced image, builds a convolutional neural network with the enhanced image as an input, and uses the convolutional neural network’s ability to classify images to recognise and classify underwater target radiation noise. The experimental results demonstrate that the training time of the method described in this paper is longer than that of the traditional machine learning method, yet the recognition and classification accuracy is significantly higher.

Pavement distress detection is crucial in road health assessment and monitoring. However, there are still some challenges in extracting pavement distress based on deep learning: such as insufficient segmentation, extraction errors and discontinuities. In this paper, we propose DARNet, a network for pavement distress extraction. A Distress Aware Attention Module (DAAM) is proposed to solve the problem of discontinuity in distress extraction due to inaccurate recovery of distress pixels during upsampling. Based on the characteristics of distress morphology, a Refinement Extraction Module (REM) is designed to effectively capture horizontal and vertical pavement damage features by fusing high-level and low-level features, which improves the accuracy of the model in extracting details of pavement damage information. Finally, a Weighted Cross-Entropy Loss function (WCEL) is introduced to assign weights according to the distance of the pixel point to the boundary of the distress, which solves the problem that the traditional cross entropy function treats each pixel point equally. We also propose a set of pavement distress datasets LNTU_RDD_GS, and the experimental results show that DARNet can reach 82.68% mIoU and 90.13% F score in the datasets in this paper, 80.63% mIoU and 88.35% F score in the four public datasets.

Pilates has established itself as a powerful instrument for a particular kind of rehabilitation therapy. Pilates can benefit people at all stages of life, especially in terms of lowering pain and disability, through enhancing physical performance and condition. It is now well established that Pilates is also likely to lead to secondary injuries if the exercise posture is not standard, so professional guidance and supervision is particularly important. This paper proposes classifying Pilates poses accurately using deep learning techniques. Ten Pilates poses are described, including Criss Cross, Rolling Ball, Spine Twist, High Plank to Pike, One Hundred, Climber, Squat, kneel Side Kick, Roll Up and Bird Dog. For each Pilates poses in this paper, 37 videos from various teachers and perspectives were gathered from the YouTube. The dataset consisted of 370 videos. The 3D human skeletal keypoints in the video frames are identified using MediaPipe in this study, and convolutional neural network(CNN) to extract features from the obtained human keypoints, combined with the Long Short-Term Memory(LSTM) algorithm to recognize Pilates poses in real time video. The model achieved an accuracy of 94.37% for the recognition of Pilates poses in single frame. An accuracy of 95.70% was achieved after polling predictions on 120 frames of video.

Steel surface defect detection is an indispensable part of industrial production and processing processes. It helps to reduce production costs, ensure product quality, improve production safety and compliance, and maintain sustainability and competitiveness. To address the low detection accuracy of traditional methods, this paper developed and investigated an improved algorithm based on YOLO for steel surface defect detection: ECA(Efficient Channel Attention) -SimSPPF (Simplified Spatial Pooling - Fast) -SIoU (Scylla Intersection over Union) -Yolov5. First, deformable convolutions were used to replace some conventional convolutions in the model, which expanded the receptive field and improved detection accuracy. Additionally, Efficient Channel Attention was integrated into the model to improve the weight of important information. Then, the SimSPPF was employed in place of the SPP module in the model, reducing computational complexity. Finally, the SIoU loss function was utilized to handle bounding box regression more effectively. The paper conducted different ablation experiments, and the improved ECA-SimSPPF-SIoU-Yolov5 algorithm demonstrated superior detection performance. Using the NEU-DET dataset, the mAP reached 78.8%, which was a 7.1% improvement higher than the original model, while the Recall reached 76.4% and improved by 3.7% compared to the original model. The improved model showed significant improvements in terms of mAP and Recall. Furthermore, the paper conducted multiple comparative experiments, comparing the model with other attention mechanisms and loss functions. The results demonstrated that the improved ECA-SimSPPF-SIoU-Yolov5 algorithm achieved good detection results in terms of mAP and Recall. In the third comparative experiment, the model was compared with YOLOv5 model with different network depths and the latest Yolov8 model, and the improved model also achieved good detection accuracy.

O2O (Online to Offline) can analyze users’ behaviors according to data mining, realize personalized marketing and improve marketing effect. The “push” delivery method of O2O coupons ignores the active participation and user experience of users, and the pertinence and effectiveness of delivery are greatly affected. Aiming at the simple structure of a single network and mainly relying on artificial construction to extract features, in order to improve the utilization efficiency of deep-seated features in the model and effectively extract multi-level features, this paper introduces Bi-GRU according to the time-series characteristics of O2O consumption behavior, and proposes a new multi-grained attention mechanism. First, build a multi-dimensional consumer behavior feature project; Secondly, using convolutional neural network (CNN) and Bi-GRU to extract local and global features; Finally, multi-level and multi-grained information is extracted by using multi-grained attention to avoid the loss of hierarchical structure information, enrich feature vectors and further improve model performance. Using real O2O coupons and data sets, the CB-MA model proposed in this paper achieves 93.29% accuracy, 91.72% AUC and 0.0332 Loss. The results show that multi-grained attention mechanism can extract multi-level features more effectively than traditional attention mechanism. At the same time, CNN and Bi-GRU are combined to learn local and global features at the same time, and the correlation information of time and space is extracted.

Recently, a novel neural architecture search method, which is referred to as DynamicNAS (Dynamic Neural Architecture Search) in this paper, has shown great potential. Not only can various sizes of models be trained with a single training session through DynamicNAS, but the subnets trained by DynamicNAS show improved performance compared to the subnets trained by conventional methods. Although DynamicNAS has many strengths compared to conventional NAS, it has the drawback that different types of operations cannot be used simultaneously within a layer as a search space. In this paper, we present a method that allows DynamicNAS to use different types of operations in a layer as a search space, without undermining the benefits of DynamicNAS, such as one-time training and superior subnet performance. Our experiments show that common operation mixing methods, such as convex combination and set sampling, are inadequate for the problem, although they have a structure that is similar to the proposed method. The proposed method finds, from a supernet of hybrid operations, a superior architecture that cannot be found from a single-operation supernet.

Deep convolutional neural networks have made some progress in skin lesion classification and cancer diagnosis, but there are still some problems to be solved, such as the challenge of small inter-class feature differences and large intra-class feature differences, which might limit the classification performance of the model as high-level and low-level features are not properly utilized. This paper proposes a multi-scale skin cancer image identification network using self-interactive attention pyramid and cross-layer bilinear-trilinear pooling(SPCB-Net), which mainly consists of three proposed sub-modules that are the self-interacting attention pyramid (SAP), the across-layer bilinear-trilinear pooling operation and the global average algorithm(GAA). The SPCB-Net is applied to two representative datasets of medical images in dermatology and histopathology (HAM10000 and NCT-CRC-HE-100K) to demonstrate the effectiveness of in the skin lesion classification. SPCB-Net(ResNet101) achieves 97.10% and 99.87% accuracy on HAM10000 and NCT-CRC-HE-100K respectively, which are both achieved performance improvements of 0.4% compared to the state-of-the-art models. In addition, a large number of experiments on HAM10000 show that the interactive attention pyramid(SPA) proposed in this paper is superior to the common attention module, and the method with a cross-layer bilinear-trilinear pooling is superior to the cross-layer trilinear pooling method. SPCB-Net is configured on Vgg19 and ResNet101 to evaluate the effectiveness of our proposed module. The experimental results show that SPCB-Net has shown state-of-the-art performance in the two field of dermatology and histopathology. Therefore, it is not only well qualified for the task of identifying skin cancer image but also has the potential to identify skin cancer by identifying pathological tissue.

Dermatological conditions are a global health concern affecting all age groups. They encompass various skin issues such as rashes, moles, acne, blisters, hives, nodules, cysts, macules, and papules. Early diagnosis of dermatological conditions is crucial to prevent skin damage and chronic diseases. Recent advancements in artificial intelligence and medical image processing, particularly through deep learning, have significantly improved the precision and efficiency of dermatological disease detection by dermatologists. This paper thoroughly examines deep learning-based methods for detecting dermatological conditions from dermoscopic images. Specifically, it presents study of 22 methods that are used to detect dermatological conditions such as basal cell carcinoma, melanocytic nevus, seborrheic keratosis, psoriasis, benign keratosis, melanoma, acne, cold sore, warts, eczema, hives, shingles, scar, skin tag, inflammatory hyper pigment, cyst, dark circle, blackhead, burn and skin rash. It also covers clinical diagnostic methods for dermatological conditions and the need for computer-aided diagnosis. In this paper, we have also proposed the categorization of deep learning-based dermatological condition detection methods. Moreover, a comprehensive summary of these methods is presented. In addition, this paper summarizes the majority of the datasets available for computer-aided detection of dermatological conditions. Furthermore, it presents enormous challenges and potential future research directions. This survey informs researchers about the latest advancements in deep learning-based detection methods for dermatological conditions and allows dermatologists to stay updated on technological breakthroughs in this area.

In smart homes, the management of energy is gaining huge significance among researchers in recent times. This paper presents a system for predicting power utilization and scheduling household appliances in smart homes. The system utilizes a combination of Grey Wolf optimization (GWO), Convolutional Neural Networks (CNN), and Long Short-Term Memory (LSTM) to improve energy management. The GWO algorithm is used to enhance the performance of the CNN-LSTM model. GWO is an optimization algorithm inspired by the hunting behaviour of grey wolves. It helps in finding optimal solutions for complex problems by mimicking the social hierarchy and hunting mechanisms of wolves. The fusion of CNN and LSTM serves as a pattern finding strategy for energy management. CNN is effective in extracting spatial features from data, while LSTM can capture temporal dependencies. By combining these two approaches, the model can analyze energy consumption patterns and make accurate predictions. To evaluate the performance of the proposed model, the paper uses three error metrics: Root Mean Square Error (RMSE), Mean Square Error (MSE), and Mean Absolute Error (MAE). The reported values of RMSE, MSE, and MAE are 0.6213, 0.3860, and 0.2808, respectively. These metrics indicate the accuracy of the model’s predictions, with lower values indicating better performance. Furthermore, this paper compares the proposed approach with the existing baseline models to access its superiority. According to the results, the proposed model outperforms the existing approaches in terms of prediction accuracy, as it achieves lower errors, compared to the baseline models. In summary, the proposed GWO-based CNN-LSTM network demonstrates improved prediction accuracy compared to the existing approaches, as indicated by the evolution metrics.

Finding relevant research papers can be a challenging task due to the enormous number of scientific publications released each year. Recently, the scientific community has been diving deep into citation analysis, specifically examining the content of papers to identify more crucial documents. Citations serve as potential parameters for establishing connections between research articles. They have been widely utilized for various academic purposes, including calculating journal impact factors, determining researchers’ h-index, allocating research grants, and pinpointing the latest research trends. However, researchers have argued that not all citations carry equal weight in terms of influence. Consequently, alternative techniques have been proposed to identify significant citations based on content, metadata, and bibliographic information. Nonetheless, the current state-of-the-art approaches still require further refinement. Additionally, the application of deep learning models and word embedding techniques in this context has not been extensively studied.In this research work, we present an approach consisting of two primary modules: 1) Section-wise citation count, and 2) metadata-based analysis of citation intent. Our study involves conducting several experiments using deep learning models in conjunction with FastText, word2vec, and BERT-based word embeddings to perform citation analysis. These experiments were carried out using two benchmark datasets, and the results were compared with a contemporary study that employed a rich set of content-based features for classification. Our findings reveal that the deep learning CNN model, coupled with FastText word embeddings, achieves the best results in terms of accuracy, precision, and recall. It outperforms the existing state-of-the-art model, achieving a precision score of 0.97

Visible and infrared image fusion (VIF) aims at remodeling an informative and panoramic image for subsequent image processing or human vision. Due to the widespread application in military and civil fields, the VIF technology has achieved considerable development in recent decades. However, the assignment of weights and the selection of fusion rules seriously restrict the performance improvement of most existing fusion algorithms. In response to this issue, an innovative and efficient VIF model based on convolutional neural network (CNN) is proposed in this paper. Firstly, multi-layer convolution kernel is performed on two source images with a multi-scale manner for extracting the salient image features. Secondly, the extracted feature maps are concatenated along the number of channels. Finally, the fusion feature maps are reconstructed to achieve the fusion images. The main innovation of this paper is to adequately preserve meaningful details and adaptively integrate features information driven by source image information in CNN learning model. In addition, in order to adequately train the network model, we generate a large-scale and high-resolution image training dataset based on COCO dataset. Compared with the existing fusion methods, experiment results indicate that the proposed method not only achieves universally outstanding visual quality and objective metrics but also has some advantages in terms of runtime efficiency compared to other neural network algorithms.

This paper proposes a technique for transforming low-resolution (LR) simulations of hair and fur into high-resolution (HR) representations without noise, using strand geometry images in the form of lines and Convolutional Neural Network (CNN or ConvNet). LR and HR data for training are obtained through physics-based simulations, and LR-HR data pairs are set up. The training data involves converting the positions of hair strands into geometry images for use in the learning process. The proposed strand network in this paper is used as an image synthesizer to upscale LR images to HR images. When the obtained HR geometry image is transformed back into HR hair through the nonlinear transform proposed in this paper, it effectively captures the elastic movement of hair that is challenging to represent with a single mapping function alone. Furthermore, we extended the solver to accommodate hair structured in a triangular mesh form. We also propose a method for adaptively refining the hair particles of strands using the Bezier curves algorithm. Additionally, we improve the algorithm to ensure that the proposed strand geometry image can work effectively even in adaptively sampled hair particle structures. As a result, the neural network enables the stable generation of strands not only for straight and curly hair but also in various environments where different external forces are applied. Furthermore, the proposed research generates results faster than previous physics-based simulations and can be easily executed without requiring a deep understanding of complex numerical analysis processes.

In the harsh electromagnetic environment with strong interference, the prior information of the received signal can not be fully obtained, and considering the complex and variable modulation modes, the modulation recognition of radio signals has brought great trouble. In this paper, we propose a method for automatic modulation recognition based on deep convolutional neural network for channel and spatial self-attention mechanism by combining the architecture of feature autonomous learning of deep learning. The correlation of input vectors in channel and spatial dimensions is enhanced by a self-attentive mechanism, and the number of layers of network structure, connection method, pooling method, and hyper-parameters are optimized, to enhance the overall fitting ability of the network and improve the accuracy and robustness of modulation recognition. On the RadioML2016.10A dataset, the proposed method is compared with the previous baseline method at different signal-to-noise ratios. The experimental results show that the performance of this paper’s method is better in the identification of 16QAM versus 64QAM.

Recognition of floor plans has been a challenging and popular task. However, existing approaches are struggling to make accurate room-level unified predictions, seriously limiting their visual quality and applicability. In this paper, we propose a novel approach to recognize the floor plan layouts with a newly proposed Offset-Guided Attention mechanism to improve the semantic consistency within a room. In addition, we present a Feature Fusion Attention module that leverages the channel-wise attention to encourage the consistency of the room, wall, and door predictions, further enhancing the room-level semantic consistency. Experimental results manifest our approach is able to improve room-level semantic consistency and outperforms the existing works both qualitatively and quantitatively. To summarize, this paper tackles inconsistent semantic prediction issues in floor plan segmentation and enhances visual plausibility, quantitative performance, and practical applicability.

Network technology has had a distinctive impact on the entire human civilization and has become an important factor of production in many countries and regions. However, with the widespread popularity of network technology, security flaws have been scattered in various fields, and potential crises may break out by attackers at any time. Therefore, it is crucial to establish a traffic monitoring mechanism for network systems. Some researchers have already implemented intrusion detection models by convolutional neural networks (CNNs) combined with attention mechanisms and achieved good results. However, few attempts have been made to improve the computational efficiency of the model by organizing the appropriate image structure, and the integration of attention mechanisms could be further enhanced. In this study, an attention-based CNN intrusion detection model has been proposed. Together with the image generation methods described in this paper, an efficient and accurate processing flow is formed. To optimize the use of the feature information in the experiments, the feature fields in the experimental images were arranged according to the results of their importance analysis. And a more integrated attention mechanism has been applied to the CNN for building the detection model. A series of comparative experiments were conducted on a subset of the CSE-CIC-IDS2018 dataset, and the results show that the detection process and model proposed in this paper can swiftly complete the detection procedure while maintaining high accuracy.

The prostate tissue structure is complex, the shape and size change is relatively large, and the surrounding anatomical structure is complex, so the task of segmenting prostate and prostate cancer is somewhat challenging. In this paper, the idea of deformable convolution is combined with the U-net algorithm widely used in medical image segmentation. By using the deformable convolution module at a specific position in the ordinary U-Net network structure, additional offsets can be added to the convolution operator and the spatial sampling position can be changed by learning the offset of the target segmentation area. The fixed receptive field of the traditional convolution operator is shifted to an adaptive receptive field that can feel the change of features, and the segmentation accuracy of the target area is improved. Experiments show that the algorithm can improve the accuracy of prostate segmentation. In this paper, the segmentation model trained with healthy prostate data is transferred to the prostate cancer data set for secondary training by simulating the way doctors read pictures. Experiments show that the segmentation effect of the lesion area is significantly improved compared with the network model trained directly with small sample prostate cancer data. The research results can provide further exploration ideas for the application of medical domain knowledge in deep learning models.

High-fidelity information, such as 4K quality videos and photographs, is increasing as high-speed internet access becomes more widespread and less expensive. Even though camera sensors’ performance is constantly improving, artificially enhanced photos and videos created by intelligent image processing algorithms have significantly improved image fidelity in recent years. Single image super-resolution is a class of algorithms that produces a high-resolution image from a given low-resolution image. Since the advent of deep learning a decade ago, this field has made significant strides. This paper presents a comprehensive review of the deep learning assisted single image super-resolution domain including generative adversarial network (GAN) models that discusses the prominent architectures, models used, and their merits and demerits. The reason behind covering the GAN models is that it is been known to perform better than the conventional deep learning methods given the resources and the time. For real-world applications with noise and other issues that can cause low-fidelity super resolution (SR) images, we examine another solution based on GAN model. This GAN model-based technique popularly known as blind super resolution is more resilient. We examined the various super-resolution techniques by varying image scaling factors (i.e., 2x, 3x, 4x) to measure PSNR and SSIM metrics for the different datasets. PSNR across the different datasets covered in the experimental Section shows an average of 14–17 % decrease in the score as we move up the image resolution scale from 2x to 4x. This is observed across all the datasets and for every model mentioned in the experimental Section of the paper. The results also show that blind super-resolution outperforms the conventional deep learning methods and the more complex GAN models. GAN models are complex and preferred when the upscale factor is high, while residual and dense models are recommended for smaller upscaling factors. This paper also discusses the applications of image super-resolution, and finally, the paper is concluded with challenges and future directions.

In order to solve the problem of low accuracy and efficiency in printed circuit board(PCB) defect detection using reference methods, a Transformer-YOLO network detection model is proposed. Firstly, an improved clustering algorithm is used to generate the anchor box suitable for the PCB defect data set of this paper. Secondly, abandoning the traditional idea of using convolutional neural network to extract image feature, Swin Transformer is used as the feature extraction network, which can effectively establish the dependency between image features. Finally, to modify the order of the channels in the feature map and enable the network to more effectively focus on the information with greater value, the convolution and attention mechanism module is added to the feature detection network component. Comparing the network model proposed in this paper with Faster R-CNN, SSD, YOLOv3, YOLOv4 and YOLOv5, the experimental results show that the proposed model improves the accuracy by 23.90%, 15.51%, 10.70%, 7.83% and 6.12% respectively, which is better than other most mainstream target detection models and has relatively small volume.

Machine learning has been applied in healthcare domain for the development of smart devices to improve the life of the elderly persons in the society. Taking care of elderly person in the society is a critical issue that need automation. To proffer solution, many researchers developed deep learning algorithms smart devices for detecting elderly behavior to improve the elderly healthcare. Despite the progress made in the applications of deep learning algorithms in elderly healthcare systems, to the best of the author’s knowledge no comprehensive recent development has been published on this interesting research area especially focusing on deep learning. In this paper, we presented a comprehensive recent development on the advances, methods and real world applications on developing smart devices for detecting elderly behavior for use in smart home, smart clinic, smart hospital and smart elderly nursing home for elderly person’s healthcare. Theories of the deep learning algorithms, recent development recorded as regard to the applicability of deep learning in elderly healthcare systems and case studies were discussed. A taxonomy based on the data extracted from the applicability of deep learning algorithms in elderly healthcare systems is created to ease pointing out areas that need more attention. The article shows that the deep learning algorithm that received tremendous attention from researchers is convolutional neural network architecture and its variants. To help in future development of the research area, we highlighted the challenges associated to the applicability of deep learning algorithms in elderly healthcare system and pointed out new point of view for future research. The research community can use our review as a benchmark for proposing novel deep learning algorithms based smart devices to detect elderly behavior for elderly healthcare systems. Industries and organizations can use the paper as a guide in selecting machine learning based smart device for detecting elderly behavior for elderly healthcare support.

Rolling element bearings are one of the important components in rotating machines. Therefore, many studies on bearing diagnosis have been conducted with artificial intelligence (AI) to do maintenance on the machines on time. In general, AI successfully diagnoses the defects of bearing when it is trained with the sufficient data of a specific machine, but it hardly provides reasonable results when it is untrained or insufficiently trained. However, it is hard to obtain sufficient data even for a specific machine in practice. In this paper, a new method was developed to increase training data by transferring the cross-domain data into the common-domain data. Therefore, all the data from different kinds of machines with various bearings can be combined as a big training data. Bearings under consideration in this paper have different specifications and characteristics. In transferring into the common-domain, it is important to get rid of structural and environmental noise by signal processing, which makes it plausible to extract common features. With the common-domain data, one-dimensional convolutional neural network (1D-CNN) with feature domain adaptation is applied and successfully classifies the defects of each bearing. Moreover, 1D-CNN combined with support vector machine (SVM) can also classify defects successfully without feature domain adaptation, which makes it possible to train the model only with normal data of the machine in concern. To verify the proposed method, not only the bearing data from Case Western Reserve University and Paderborn University but also the bearing data with flow noise of Ajou University are used.

This paper addresses the problem of super-resolution of environment matting of transparent objects. In contrast to traditional methods of environment matting of transparent objects, which often require a large number of input images or complex camera setups, recent approaches using convolutional neural networks are more practical. In particular, after training, they can generate the environment mattes using a single image. However, they still do not have super-resolution capabilities. This paper first proposes an encoder-decoder network with restoration units for super-resolution environment matting, called Enhanced Transparent Object Matting Network (ETOM-Net). Then, we introduce a refinement phase to improve the details of the output further. The ETOM-Net effectively recovers lost features in the LR input images and produces visually plausible HR environment mattes and the corresponding reconstructed images, demonstrating our method’s effectiveness.

In this paper, we apply a genetic algorithm (GA) for feature selection, wrapper approach, on wavelet scattering (WS) second-order coefficients to reduce the large frequency dimension (>500). The evaluation demonstrates the capability of GA to reduce the dimension space by approximately 30% while ensuring a minimum performance drop. Furthermore, the reduced WS directly impacts the training time of the convolutional neural network, by reducing the computational time by 20% to 32%. The paper extends its scopes to explore GA for feature selection on multiple timescales of WS: 46ms, 92ms, 185ms, and 371ms. Incorporating multiple timescales has improved classification performance (by ~ around 2.5%) as an acoustic representation usually contains information at different time scales. However, it can increase computational cost due to the larger frequency dimension of 1851. With the application of GA for feature selection, the frequency dimension is reduced by 50%, saving around 40% computational time, thus increasing the classification performance by 3% compared to a vanilla WS. Lastly, the entire implementations are evaluated using the Detection and Classification of Acoustic Scenes and Events (DCASE) 2020 dataset, and the proposed multiple timescales model achieves 73.32% of classification accuracy.

The key towards a low complexity model for convolution neural network is in controlling the number of parameters of the network and ensuring that the input representation is not extremely large. Hence, to tackle low complexity for acoustic scene classification (ASC), this paper proposed an enhanced wavelet scattering representation with a combination of mobile network modules and shuffling modules. While wavelet scattering comprises wavelet transform with multiple wavelet scales, the averaging operation to make the wavelet scattering invariant to translation limit the maximum timescale. Hence, wavelet scattering is affected by Heisenberg’s Uncertainty Principle. However, creating an input representation with multiple timescales does not meet the brief of low complexity modelling. Hence, we proposed a simple mixing of the first and second order with different timescales. The result is an input representation with nearly the same dimension as the usual wavelet scattering but with enhanced multiscale. To further leverage the ‘interleaved’ wavelet scattering, this paper presents sub-spectral shuffling inspired by shuffling modules that use stochasticity to improve the model’s generalization. Unlike channel shuffle that shuffles channel-wise and spatial shuffle that shuffles pixel-wise, sub-spectral shuffle aims at shuffling the feature maps frequency-wise with the concept of binning. Each bin is shuffled, so the high-frequency spectrum is shuffled to low-frequency spectrum position. As such, the model learns the general acoustic profile of a scene rather than memorizing what is happening at the low-frequency or high-frequency spectrum is erratic for ASC. In addition, this paper also studied temporal shuffling, which shuffles the feature maps temporal-wise, and evaluated sub-spectral shuffling, temporal shuffling, and channel shuffling individually. Our results demonstrated the superiority of sub-spectral shuffling and the modularity of shuffling modules. We then evaluate various combinations of the three shuffling modules on three acoustic scene classification datasets. Our best model combines the three shuffling modules and achieves 70.6% classification accuracy on DCASE 2021 Task 1a dataset, 82.15% on ESC-50 dataset, 81% on Urbansound8K, with ~65K parameters and a size of 126.6KB. In addition, the inclusion of shuffling modules has increased the model performance. Sub-spectral shuffling is especially useful in improving logloss, a metric used to determine the confidence level of the model.

The rise of Telemedicine has revolutionized how patients are being treated, leading to several advantages such as enhanced health analysis tools, accessible remote healthcare, basic diagnostic of health parameters, etc. The advent of the Internet of Things (IoT), Artificial Intelligence (AI) and their incorporation into Telemedicine extends the potential of health benefits of Telemedicine even further. Therefore, the synergy between AI, IoT, and Telemedicine creates diverse innovative scenarios for integrating cyber-physical systems into medical health to provide remote monitoring and interactive assistance to patients. Data from World Health Organization reports that 7.4 million people died because of Atrial Fibrillation (AF), recognizing the most common arrhythmia associated with human heart rate. Causes like unhealthy diet, smoking, poor resources to go to the doctor and based on research studies, about 12 and 17.9 million of people will be suffering the AF in the USA and Europe, in 2050 and 2060, respectively. The AF as a cardiovascular disease is becoming an important public health issue to tackle. By using a systematic approach, this paper reviews recent contributions related to the acquisition of heart beats, arrhythmia detection, IoT, and visualization. In particular, by analysing the most closely related papers on Convolutional Neural Network (CNN) and IoT devices in heart disease diagnostics, we present a summary of the main research gaps with suggested directions for future research.

Deep learning-based land cover and land use (LCLU) classification systems are a significant aspiration for remote sensing communities. In nature, remote sensing images have various properties that need to be analyzed. Analyzing and interpreting image properties is difficult due to the nature of the image, the sensor technology’s capability, and other determinant variables such as seasons and weather conditions. The problem is essential for environmental monitoring, agricultural decision-making, and urban planning if it can be supported by deep learning systems. Therefore, deep learning approaches are proposed to quickly analyze and interpret the remote sensing image to classify the LCLU. The deep learning methods could be designed starting from scratch or using pre-trained networks. However, there are few comparisons of deep learning methods developed from scratch and trained on pre-trained networks. Thus, we proposed evaluating and comparing the deep learning models convolutional neural network feature extractor (CNN-FE) by developing it from scratch, transfer learning, and fine-tuning it for the LCLU classification system using remote sensed images. Using CNN-FE, TL, and fine-tuning deep learning models as examples, this paper compares and analyzes deep learning algorithms for remote sensed image classification. After developing and training each deep learning model on the UCM dataset, we evaluated and compared their performances using the performance measurement metrics accuracy, precision, recall, f1-score, and confusion matrix. The proposed deep learning algorithms can adapt and learn the features of the remote sensing images, and the TL and fine-tuning classification performances are significantly improved. As a result of the efficient time used for training the models, this paper discovered that the fine-tuned deep learning model achieved profound accuracy performance results in the UCM dataset.

Network pruning can effectively reduce a model's capacity and computational load, thereby making model deployment in mobile devices less difficult than that without pruning. To improve the pruning rate of the model while maintaining the kernel's feature extraction ability, this paper designs a progressive kernel pruning method for CNN model compression based on the proposed information mapping sparsity index. This method first prunes the kernels in the filter and then prunes the kernels in the convolution layer when the model reaches a certain compression ratio. The whole process is called progressive kernel pruning (PKP). For the kernel pruning process, this paper defines the information mapping sparse index (IMSI), which is used to measure the mapping ability of the convolution kernel related to the amount of information transferred by the convolution operation. When pruning the kernels of filters and layers, according to the IMSI, the kernels with the strongest mapping abilities are retained to transfer as much information as possible with the least number of kernels. Progressive kernel pruning can make use of the characteristic that the model is easy to optimize when kernel pruning in the filter, and it avoids having the model easily fall into local optima when kernel pruning in the layer directly. The experimental results on the CIFAR-10/100 and ImageNet datasets show that compared to the existing CNN model compression methods, the IMSI-based progressive kernel pruning method exhibits better compression performance in processing the model compression tasks that are currently popular. In particular, pruning VGG-16 on CIFAR-10 with our model achieves a compression ratio of 80.8x and an acceleration ratio of 14.8×, which are 5.8× and 4.2× higher than the best results at present, respectively, and the classification accuracy decreases by only 0.59% relative to that of the baseline.

The airway tree is one of the most important part in human respiratory system. Airway segmentation plays a crucial role in pulmonary disease diagnosis, localization and surgical navigation. We propose a novel method to improve airway segmentation in thoracic computed tomography(CT) using deep learning. In order to take into account the multi-scale changes of the airway and achieve accurate airway segmentation, we design an end-to-end Tiny Atrous Convolutional Network (TACNet) based on 3D convolution neural network. In view of the difficulty of classification due to the numerous branches of airway, two evaluation factors, namely, the angle of airway bifurcation and the buffer length of airway bifurcation are designed, which are used for airway classification by combining the centerline extraction. We train TACNet on the inspirator thoracic CT scans with ground truth, which are generated by clinicians and evaluate on our own clinical data sets and EXACT’09 data sets. Compared with state-of-the-art airway segmentation algorithms, proposed algorithm in this paper achieve very competitive results in 20 test datasets of the EXACT’09 challenge. The experimental results show that the algorithm proposed in this paper has high robustness and advantages regardless of airway segmentation or airway classification. In 20 test datasets of the EXACT’09 challenge, the average airway tree length detection rate is the best in the public literatures.

In the advanced television systems committee (ATSC) 3.0 system, the concept of flexibility is significant for supporting backward compatibility within the same ATSC 3.0 system. However, since the conventional bootstrap signal detection scheme is difficult to support the flexibility, the conventional bootstrap signal detection scheme should be newly designed according to the change of version. In this paper, a convolution neural network (CNN) model for bootstrap signal detection in ATSC 3.0 is proposed to maintain the flexibility of bootstrap. Additionally, for minimizing the loss of error performance of CNN-based bootstrap detection scheme, this paper proposes two dimensional alternate array to utilize the correlation of the adjacent bootstrap symbol and proposes the offline learning method using the bootstrap signal corrupted by noise to improve the error performance.

In recent years, researchers have proposed many deep learning (DL) methods for various tasks, and particularly face recognition (FR) made an enormous leap using these techniques. Deep FR systems benefit from the hierarchical architecture of the DL methods to learn discriminative face representation. Therefore, DL techniques significantly improve state-of-the-art performance on FR systems and encourage diverse and efficient real-world applications. In this paper, we present a comprehensive analysis of various FR systems that leverage the different types of DL techniques, and for the study, we summarize 171 recent contributions from this area. We discuss the papers related to different algorithms, architectures, loss functions, activation functions, datasets, challenges, improvement ideas, current and future trends of DL-based FR systems. We provide a detailed discussion of various DL methods to understand the current state-of-the-art, and then we discuss various activation and loss functions for the methods. Additionally, we summarize different datasets used widely for FR tasks and discuss challenges related to illumination, expression, pose variations, and occlusion. Finally, we discuss improvement ideas, current and future trends of FR tasks.

Recently, there is a rapid and wide increase in fake news, defined as provably incorrect information spread with the goal of fraud. The spread of this type of misinformation is a severe danger to social cohesiveness and well-being since it increases political polarisation and people’s distrust of their leaders. Thus, fake news is a phenomenon that is having a significant impact on our social lives, particularly in politics. This paper proposes novel approaches based on Machine Learning (ML) and Deep Learning (DL) for the fake news detection system to address this phenomenon. The main aim of this paper is to find the optimal model that obtains high accuracy performance. Therefore, we propose an optimized Convolutional Neural Network model to detect fake news (OPCNN-FAKE). We compare the performance of the OPCNN-FAKE with Recurrent Neural Network (RNN), Long Short-Term Memory (LSTM), and The six regular ML techniques: Decision Tree (DT), logistic Regression (LR), K Nearest Neighbor (KNN), Random Forest (RF), Support Vector Machine (SVM), and Naive Bayes (NB) using four fake news benchmark datasets. Grid search and hyperopt optimization techniques have been used to optimize the parameters of ML and DL, respectively. In addition, N-gram and Term Frequency—Inverse Document Frequency (TF-IDF) have been used to extract features from the benchmark datasets for regular ML, while Glove word embedding has been used to represent features as a feature matrix for DL models. To evaluate the performance of the OPCNN-FAKE, accuracy, precision, recall, F1-measure were applied to validate the results. The results show that OPCNN-FAKE model has achieved the best performance for each dataset compared with other models. Furthermore, the OPCNN-FAKE has a higher performance of cross-validation results and testing results over the other models, which indicates that the OPCNN-FAKE for fake news detection is significantly better than the other models.

Automated human pose estimation is evolving as an exciting research area in human activity detection. It includes sophisticated applications such as malpractice detection in the examination, distracted driving, gesture detection, etc., and requires robust and reliable pose estimation techniques. These applications help to map the attention of the user with head pose estimation (HPE) metrics supported by emotion and gaze analysis. This paper solves the problem of attention score estimation with HPE. The proposed method ensures ease of implementation while addressing head pose estimation using 68 facial features. Further, to attain reliability and precision, head pose estimation has been implemented as a regression task. The coordinate pair angle method (CPAM) with deep neural network (DNN) regression and elastic net regression is carried out. The use of DNN ensures precision on low lighting, distorted or occluded images. CPAM methodology leverages facial landmark detection and angular difference to estimate head pose. Experimentation results showed that the proposed model could handle large datasets, real-time data processing, significant pose variations, partial occlusions, and diverse facial expressions with a mean absolute error (MAE) of 3° and less. The proposed system was evaluated on three standard databases: the 300W across large poses (300W-LP) dataset, annotated facial landmarks in the wild (AFLW2000) dataset, and the national institute of mental health child emotional faces picture set (NIMH-ChEFS) dataset. The results achieved are on par with recent state-of-the-art methodologies such as anisotropic angle distribution learning (AADL), joint head pose estimation and face alignment algorithm (JFA), rotation axis focused attention network (RAFA-Net), and propose an MAE ranging up to 6°. The paper could achieve remarkable results for attention span prediction using head pose estimation and for many possible future applications.

This paper presents drone classification at millimeter-wave (mmWave) radars using the deep learning (DL) technique. The adoption mmWave technology in radar systems enables better resolution and aid in detecting smaller drones. Using radar cross-section (RCS) signature enables us to detect malicious drones and suitable action can be taken by respective authorities. Existing drone classification converts the RCS signature into images and then performs drone classification using a convolution neural network (CNN). Converting every signature into an image induces additional computation overhead; further CNN model is trained considering fixed learning rate. Thus, when using CNN-based drone classification under a highly dynamic environment exhibit poor classification accuracy. This paper present an improved long short-term memory (LSTM) by introducing a weight optimization model that can reduce computation overhead by not allowing the gradient to not flow through hidden states of the LSTM model. Further, present adaptive learning rate optimizing (ALRO) model for training the LSTM model. Experiment outcome shows LSTM-ALRO achieves much better drone detection accuracies of 99.88% when compared with the existing CNN-based drone classification model.

Existing text sentiment analysis methods mostly rely on a large number of language knowledge and sentiment resources. This paper proposes the Multi-channel convolution and bidirectional GRU multi-head attention capsule (AT-MC-BiGRU-Capsule), which uses vector neurons to replace scalar neurons to model text emotions, and uses capsules to characterize text emotions. In addition, traditional methods cannot extract the multi-level features of text sequence well. Multi-head attention can encode the dependencies between words, capture sentiment words in text, and using Convolutional Neural Network (CNN) and Bidirectional gated recurrent unit network (Bi-GRU) to extract local features and global semantic features of text respectively, the global average pooling layer is introduced to obtain the multi-level feature representation of the text sequence more comprehensively. This paper selects three English datasets and one Chinese dataset in the general corpus of sentiment classification to conduct experiments, and achieves better results than other baseline models.

Artificial intelligence-assisted diagnosis systems are developing rapidly, but doctors are currently less aware of artificial intelligence-assisted diagnosis systems. Understanding how to allow doctors to accept and use artificial intelligence medical assistant diagnosis system can promote the implementation of artificial intelligence medical assistant diagnosis system applications. Taking into account the current difficulties faced by the Internet of Things medical consultation services, this paper proposes a business operation model based on multi-party participation and sharing of medical consultation resources. We designed the information flow, overall logic and service implementation process of the service model, and completed the construction of the artificial intelligence medical service service model. We combine IoT technology to build a vital signs monitoring environment and clarify how to use IoT devices. In the error backpropagation algorithm, there is no significant difference in the contribution of different samples to the weight change, which makes the adjustment of network parameters not easily affected by difficult medical consultation samples, thereby weakening the effect of network medical consultation. In order to solve this problem, this article defines the degree to which the sample belongs to its correct category as the confidence of medical inquiry, and divides the training sample into a dangerous sample and a safe sample according to a dynamic threshold. Based on the difficulty of medical inquiry, an improved new learning algorithm is proposed. The algorithm penalizes the loss of dangerous samples, so that the convolutional neural network pays more attention to dangerous samples and can learn more effective information. Aiming at the eight physiological characteristics of data, this paper adjusts the structure of the convolutional neural network to make it take into account the richness of data characteristics and the dynamics of data changes over time. The realized CNN optimization algorithm model has improved the prediction effect, and the accuracy rate of medical consultation reaches 90.15%, which is better than other machine learning algorithms.

Image super-resolution reconstruction has been widely used in remote sensing, medicine and other fields. In recent years, due to the rise of deep learning research and the successful application of convolutional neural networks in the image field, the super-resolution reconstruction technology based on deep learning has also achieved great development. However, there are still some problems that need to be solved. For example, the current mainstream image super-resolution algorithms based on single or multiple frames pursue high performance indicators such as PSNR and SSIM, while the reconstructed image is relatively smooth and lacks many high-frequency details. It is not conducive to application in a real environment. To address such problem, this paper proposes a super-resolution reconstruction model of sequential images based on Generative Adversarial Networks (GAN). The proposed approach combines the registration module to fuse adjacent frames, effectively use the detailed information in multiple consecutive frames, and enhances the spatio-temporality of low-resolution images in sequential images. While the GAN was used to improve the effect of image high-frequency texture detail reconstruction, WGAN was introduced to optimize model training. The reconstruction results not only improved the PSNR and SSIM indexes but also reconstructed more high-frequency detail textures. Finally, in order to further improve the perception effect, an additional registration loss item RLT is introduced in the GAN network perception loss. Through extensive experiments, it shows that the model proposed in this paper effectively obtains the information between the sequence images. When the PSNR and SSIM indicators are improve, it can reconstruct better high-frequency texture details than the current advanced multi-frame algorithms.

Global horizontal irradiance (GHI) is a critical index to indicate the output power of the photovaltaic (PV). In traditional approaches, the local GHI can be measured with very expensive instruments, and the large-area GHI collection depends on complex satellite-based models, solargis algorithms, and the high-performance computers (HPC). In this paper, a novel approach is proposed to capture the GHI conveniently and accurately. Considering the nonstationary property of the GHI on cloudy days, the GHI capturing is cast as an image regression problem. In traditional approaches, the image regression problem is treated as two parts, feature extraction (for the images) and regression model (for the regression targets), which are optimized separately and blocked the interconnections. Considering the nonlinear regression capability, a convolutional neural network (CNN) based image regression approach is proposed to provide an End-to-End solution for the cloudy day GHI capturing problem in this paper. The multilayer CNN is based on the AlexNet and VGG. The L2 (least square errors) with regularization is used as the loss function in the regression layer. For data cleaning, the Gaussian mixture model with Bayesian inference is employed to detect and eliminate the anomaly data in a nonparametric manner. The purified data are used as input data for the proposed image regression approach. In the experiments, three-month sky images and GHI data (with 1-min resolution) are provided by the National Renewable Energy Laboratory (NREL) with the HPC system. The numerical results demonstrate the feasibility and effectiveness of the proposed approach.

This paper mainly studies the problem of image-text matching in order to make image and text better match. Existing cross-media retrieval methods only make use of the information of image and part of text, that is, matching the whole image with the whole sentence, or matching some image areas with some words. In order to better reveal the potential connection between image and text semantics, this paper proposes a fusion of two levels of similarity across media images-text retrieval method, constructed the cross-media two-level network to explore the better matching between images and texts, it contains two subnets for dealing with global features and local characteristics. Specifically, in this method, the image is divided into the whole picture and some image area, the text is divided into the whole sentences and words, to study respectively, to explore the full potential alignment of images and text, and then use a two-level alignment framework is used to promote each other, fusion of two kinds of similarity can learn to complete representation of cross-media retrieval. Through the experimental evaluation on Flickr30K and MS-COCO datasets, the results show that the method in this paper can make the semantic matching of image and text more accurate, and is superior to the international popular cross-media retrieval method in various evaluation indexes.

In the textile and apparel industry, it remains a challenging task to evaluate the fabric smoothness appearance objectively. In existing studies, with computer vision technology, researchers use the hand-crafted image features and deep convolutional neural network (CNN) based image features to describe the fabric smoothness appearance. This paper presents an image classification framework to evaluate the fabric smoothness appearance degree. The framework contains a feature fusion module to fuse the hand-crafted and CNN features to take both advantages of them. The framework uses the multi-scale spatial masking model and a pre-trained CNN to extract hand-crafted and CNN features of fabric images respectively. In addition, a mislabeled sample filtering module is set in the framework, which helps to avoid the negative impact of mislabeled samples in training. In the experiments, the proposed framework achieves 85.2%, 96.1%, and 100% average evaluation accuracies under errors of 0 degree, 0.5 degree, and 1 degree respectively. The experiments on the feature fusion and mislabeled sample filtering verified their effectiveness in improving the evaluation accuracies and the label noise robustness. The proposed method outperforms the state-of-the-art methods for fabric smoothness assessment. Promisingly, this paper can provide novel research ideas for image-based fabric smoothness assessment and other similar tasks.

With the advancements in machine and deep learning algorithms, the envision of various critical real-life applications in computer vision becomes possible. One of the applications is facial sentiment analysis. Deep learning has made facial expression recognition the most trending research fields in computer vision area. Recently, deep learning-based FER models have suffered from various technological issues like under-fitting or over-fitting. It is due to either insufficient training and expression data. Motivated from the above facts, this paper presents a systematic and comprehensive survey on current state-of-art Artificial Intelligence techniques (datasets and algorithms) that provide a solution to the aforementioned issues. It also presents a taxonomy of existing facial sentiment analysis strategies in brief. Then, this paper reviews the existing novel machine and deep learning networks proposed by researchers that are specifically designed for facial expression recognition based on static images and present their merits and demerits and summarized their approach. Finally, this paper also presents the open issues and research challenges for the design of a robust facial expression recognition system.

Object tracking is a key technology in the field of intelligent transportation. To solve the partial occlusion problem in vehicle tracking, this paper analyzes the characteristics of a VGG convolutional neural network by experimental observation and describes its characteristics: (1) feature maps can be used for positioning, but they have redundancies, and (2) different layers of feature maps have different characteristics. After these characteristics are applied to vehicle tracking, a vehicle tracking algorithm is designed. For a given target vehicle, feature maps are generated on convolutional layers conv4_4 and conv5_4 of the VGG network, and the feature maps most relevant to the target vehicle are selected. These feature maps are used to capture target vehicle information and distinguish the target vehicle from backgrounds with similar appearances. The experiments use vehicle data from the LaSOT, VOT2017 and OTB2015 datasets to compare the vehicle tracking results of our proposed algorithm with those of other algorithms. The results show that the method proposed in this paper has certain advantages. According to algorithm implementation and vehicle tracking experiments, the proposed vehicle tracking method can solve the drift problem and is better than the traditional method at addressing drift problems.

This paper addresses road anomaly detection by formulating it as a classification problem and applying deep learning approaches to solve it. Besides conventional road anomalies, additional ones are introduced from the perspective of a vehicle. In order to facilitate the learning process, the paper pays a close attention to pattern representation, and proposes three sets of numeric features for representing road conditions. Also, three deep learning approaches, i.e. Deep Feedforward Network (DFN), Convolutional Neural Network (CNN), and Recurrent Neural Network (RNN), are considered to tackle the classification problem. The detectors, with respect to the three deep learning approaches, are trained and evaluated through data collected from a test vehicle driven on various road anomaly conditions. The comparison study on the detection performances is conducted by setting key hyper-parameters to certain sets of fixed values. Also, the comparison study on performances of each detector with respect to different pattern representations is conducted. The results have shown the effectiveness of the proposed approaches and the efficiency of the proposed feature representations in road anomaly detection.

Deep Convolutional Neural Networks (CNNs) can be overly dependent on training data, causing a generalization problem in which trained models may not predict real-world datasets. To address this problem, various regularization methods such as image manipulation and feature map regularization have been proposed for their strong generalization ability. In this paper, we propose a regularization method that applies both image manipulation and feature map regularization based on patches. The method proposed in this paper has a regularization effect in two stages, which makes it possible to better generalize the model. Consequently, it improves the performance of the model. Moreover, our method adds features extracted from other images in the hidden state stage, which not only makes the model robust to noise but also captures the distribution of each label. Through experiments, we show that our method performs competently on models that generate a large number of parameter and multiple feature maps for the CIFAR and Tiny-ImageNet datasets.

Breast cancer is one of the most dangerous diseases and the second largest cause of female cancer death. Breast cancer starts when malignant, cancerous lumps start to grow from the breast cells. Self-tests and Periodic clinical checks help to early diagnosis and thereby improve the survival chances significantly. The breast cancer classification is a medical method that provides researchers and scientists with a great challenge. Neural networks have recently become a popular tool in cancer data classification. In this paper, Deep Learning assisted Efficient Adaboost Algorithm (DLA-EABA) for breast cancer detection has been mathematically proposed with advanced computational techniques. In addition to traditional computer vision approaches, tumor classification methods using transfers are being actively developed through the use of deep convolutional neural networks (CNNs). This study starts with examining the CNN-based transfer learning to characterize breast masses for different diagnostic, predictive tasks or prognostic or in several imaging modalities, such as Magnetic Resonance Imaging (MRI), Ultrasound (US), digital breast tomosynthesis and mammography. The deep learning framework contains several convolutional layers, LSTM, Max-pooling layers. The classification and error estimation that has been included in a fully connected layer and a softmax layer. This paper focuses on combining these machine learning approaches with the methods of selecting features and extracting them through evaluating their output using classification and segmentation techniques to find the most appropriate approach. The experimental results show that the high accuracy level of 97.2%, Sensitivity 98.3%, and Specificity 96.5% has been compared to other existing systems.

Behavior is the first visible change in an animal species after exposure to its own or environmental stressors and is a sensitive indicator. Fish are social animals, and the abnormality of group behavior is more indicative about a particular event than individual behavior, providing more effective informeqation about environmental or group social changes. The group behavior is not only reflected in the spatial distribution, but also reflected in the temporal behavior of the group and individual movement changes under the influence of pressure factors. This paper proposes a group behavior discrimination method based on convolutional neural network and spatiotemporal information fusion, which intends to make use of the prominent performance of convolutional neural network in image recognition and state classification, and imitating the attentional mechanism of ventral channel and dorsal channel when the human brain processes visual signals. Some pressure environments are made in laboratory, the behavior states of fish shoals are recorded, and the sample database of shoals' behavior state is established by combining the spatial information of shoals' spatial distribution with the time information reflected in the movement behavior. A simple convolutional neural network is constructed to quickly identify the behavior state of fish shoals. The effects of bath size and training epoch on network training speed and recognition accuracy are discussed, and the visualization of the intermediate data of the convolutional neural network is studied. Shown from the results of experiments of this paper, different behavior states of fish shoals can be recognized and classified effectively by using the simple convolutional neural network and spatiotemporal fusion images. What's more, from the visualization of network intermediate data, it is found that the convolutional neural network has a higher discrimination power to the image edge feature than the image gray-value feature.

As an algorithm with excellent performance, convolutional neural network has been widely used in the field of image processing and achieved good results by relying on its own local receptive fields, weight sharing, pooling, and sparse connections. In order to improve the convergence speed and recognition accuracy of the convolutional neural network algorithm, this paper proposes a new convolutional neural network algorithm. First, a recurrent neural network is introduced into the convolutional neural network, and the deep features of the image are learned in parallel using the convolutional neural network and the recurrent neural network. Secondly, according to the idea of ResNet's skip convolution layer, a new residual module ShortCut3-ResNet is constructed. Then, a dual optimization model is established to realize the integrated optimization of the convolution and full connection process. Finally, the effects of various parameters of the convolutional neural network on the network performance are analyzed through simulation experiments, and the optimal network parameters of the convolutional neural network are finally set. Experimental results show that the convolutional neural network algorithm proposed in this paper can learn the diverse features of the image, and improve the accuracy of feature extraction and image recognition ability of the convolutional neural network.

It is of great importance to construct a convolutional neural network architecture in the frequency domain to explore the theory of deep learning in the frequency domain. However, due to the complexity of the construction mechanism of the forward and backward pipelines needed to train the convolutional neural network in the frequency domain, higher requirements are put forward for the representation strategy of the frequency-domain activation function and pooling method in the forward and backward pipelines. Therefore, to construct a full frequency-domain convolutional neural network architecture, it is necessary to construct a frequency-domain representation strategy with a high classification accuracy and excellent time performance. In this paper, based on a chunk decomposition mechanism and the construction principle of the frequency-domain unsaturated activation function, a frequency-domain convolutional neural network architecture is proposed. Two important representation strategies are introduced into the frequency-domain forward/backward pipeline: a frequency-domain randomized offset rectified linear unit and a frequency-domain chunk max pooling method. The former can alleviate the vanishing and exploding gradient phenomena in the frequency-domain forward/backward pipeline and ensure the convergence of the convolutional neural network architecture in the frequency-domain training stage; the latter can capture the partial location information and characteristic strength of the frequency-domain neurons and improve the classification performance of the convolutional neural network in the frequency domain. This full frequency-domain convolutional neural network architecture improves the training accuracy of the convolutional neural network in the frequency-domain pipeline. The results show that on the basis of ResNet-50 as the backbone framework, an NVIDIA GeForce CUDA(Compute Unified Device Architecture) as the training pipeline, and $4\times 4$ as the activation block size of the third-level output neuron's characteristic parameter matrix, the convolutional neural network architecture proposed in this paper can lower the top-1 error from 24.90% to 17.95%, the top-5 error from 12.85% to 9.23%. Furthermore, when the batch size is equal to 128 (in the worst-case bandwidth usage scenario), the acceleration ratio of the proposed architecture can still reach 13.0375 by selecting cuDNN as the reference model. Under the same backbone framework, the proposed architecture is tested on the MetData-1 dataset, and the classification accuracy can reach the maximum value; that is, the average difference is merely 0.18. This finding shows that the proposed architecture can improve the accuracy of the deep learning-based frequency-domain convolutional neural network model without reducing the time performance and expand the frequency-domain representation strategy of the frequency-domain activation function and pooling method.

This paper presents a deep end-to-end network for high dynamic range (HDR) imaging of dynamic scenes with background and foreground motions. Generating an HDR image from a sequence of multi-exposure images is a challenging process when the images have misalignments by being taken in a dynamic situation. Hence, recent methods first align the multi-exposure images to the reference by using patch matching, optical flow, homography transformation, or attention module before the merging. In this paper, we propose a deep network that synthesizes the aligned images as a result of blending the information from multi-exposure images, because explicitly aligning photos with different exposures is inherently a difficult problem. Specifically, the proposed network generates under/over-exposure images that are structurally aligned to the reference, by blending all the information from the dynamic multi-exposure images. Our primary idea is that blending two images in the deep-feature-domain is effective for synthesizing multi-exposure images that are structurally aligned to the reference, resulting in better-aligned images than the pixel-domain blending or geometric transformation methods. Specifically, our alignment network consists of a two-way encoder for extracting features from two images separately, several convolution layers for blending deep features, and a decoder for constructing the aligned images. The proposed network is shown to generate the aligned images with a wide range of exposure differences very well and thus can be effectively used for the HDR imaging of dynamic scenes. Moreover, by adding a simple merging network after the alignment network and training the overall system end-to-end, we obtain a performance gain compared to the recent state-of-the-art methods.

In order to solve the scattering degradation by turbulence and suspended particles in underwater imaging, traditional processing methods including image enhancement, restoration and reconstruction have been continuously researched. But most of them rely on degradation models, and there exist problems of ill-posed. Image super resolution reconstruction based on deep learning has become a hot topic in recent years. In order to further improve the effectiveness and efficiency of deep learning based methods, an improved image super-resolution reconstruction algorithm based on deep convolutional neural network is proposed in this paper. The wavelet basis which can effectively simulate the waveform and characteristics of underwater turbulence is selected to replace the neuron fitting function in order to improve the accuracy and efficiency of the algorithm. An improved dense block structure (IDB) is introduced into the network which can effectively solve the gradient disappearance problem of deep convolutional neural network and improve the training speed at the same time. The method proposed in this paper has been verified in laboratory flume, public data set and real water body. The experimental results show that under the same conditions, the proposed algorithm shows improvements on various evaluation parameters compared with DRFN, VDSR and DRCN method. So it can be concluded that the proposed method can effectively improve the quality of deep learning based reconstruction for imaging in natural water.