Search

Even with the most cutting-edge tools, treating and monitoring patients—including children, elders, and suspected COVID-19 patients—remains a challenging activity. This study aimed to track multiple COVID-19-related vital indicators using a wearable monitoring device with an Internet of Things (IOT) focus. Additionally, the technology automatically alerts the appropriate medical authorities about any breaches of confinement for potentially contagious patients by tracking patients’ real-time GPS data. The wearable sensor is connected to a network edge in the Internet of Things cloud, where data are processed and analyzed to ascertain the state of body function. The proposed system is built with three tiers of functionalities: a cloud layer using an Application Peripheral Interface (API) for mobile devices, a layer of wearable IOT sensors, and a layer of Android web for mobile devices. Each layer performs a certain purpose. Data from the IoT perception layer are initially collected in order to identify the ailments. The following layer is used to store the information in the cloud database for preventative actions, notifications, and quick reactions. The Android mobile application layer notifies and alerts the families of the potentially impacted patients. In order to recognize human activities, this work suggests a novel integrated deep neural network model called CNN-UUGRU which mixes convolutional and updated gated recurrent subunits. The efficiency of this model, which was successfully evaluated on the Kaggle dataset, is significantly higher than that of other cutting-edge deep neural models and it surpassed existing products in local and public datasets, achieving accuracy of 97.7%, precision of 96.8%, and an F-measure of 97.75%.

Accurate segmentation of brain tumors is an essential stage in treatment planning. Fully convolutional neural networks, specifically the encoder-decoder architectures such as U-net, have proven successful in medical image segmentation. However, segmenting brain tumors with complex structure requires building a deeper and wider model which increases the computational complexity and may also cause the gradient vanishing problem. Therefore, in this work, we propose a novel encoder-decoder architecture, called Inception Residual Dense Nested U-Net (IRDNU-Net). In this model carefully designed Residual and Inception modules are used in place of standard U-Net convolutional layers to increase the width of the model without increasing the computational complexity. Additionally, in the proposed architecture, the encoder and decoder are connected via a sequence of Inception-Residual densely nested paths to extract more information and increase the depth of the network while reducing the number of network parameters. The proposed segmentation architecture was evaluated on two large brain tumor segmentation benchmark datasets; the BraTS'2019 and BraTS'2020. It achieved a mean Dice similarity coefficient of 0.888 for the whole tumor region, 0.876 for the core region, and 0.819 for the enhancement region. Experimental results illuminate that IRDNU-Net outperforms U-Net by 1.8%, 11.4%, and 11.7% in the whole tumor, core tumor, and enhancing tumor, respectively. Moreover, the IRDNU-Net enables a great improvement on the accuracy compared to comparative approaches, and its ability in the face of challenging problems, such as small tumor regions, with fewer parameters. [ABSTRACT FROM AUTHOR]

With the development of convolutional neural networks, aiming at the problem of low efficiency and low accuracy in the process of wood species recognition, a recognition method using an improved convolutional neural network is proposed in this article. First, a large-scale wood dataset was constructed based on the WOOD-AUTH dataset and the data collected. Then, a new model named W_IMCNN was constructed based on Inception and mobilenetV3 networks for wood species identification. Experimental results showed that compared with other models, the proposed model had better recognition performance, such as shorter training time and higher recognition accuracy. In the data set constructed by us, the accuracy of the test set reaches 96.4%. We used WOOD-AUTH dataset to evaluate the model, and the recognition accuracy reached 98.8%. Compared with state-of-the-art methods, the effectiveness of the W_IMCNN were confirmed. [ABSTRACT FROM AUTHOR]

Mammograms have been acknowledged as one of the most reliable screening tools as well as a key diagnostic mechanism for early breast cancer detection. Though mammography is a valuable screening tool for detecting malignant growth in breasts, its competence as a diagnostic tool is heavily reliant on the radiologists' understanding. Automated systems are now widely used for detection of breast cancer. Image processing techniques were widely used in automated systems for classifying mammograms. Of late with the advent of deep learning (DL) where images can be processed directly for classification, the DL is widely researched for medical image classification. Basically, DL techniques are representationlearning methods which aid in understanding data like sounds, images as well as texts. DL algorithms have the ability to learn multiple levels of representation as well as abstraction. Residual network (ResNet) is given due consideration as a kind of highly advanced Convolutional Neural Networks (CNNs). This work has offered a potential application of Visual Geometry Group (VGG), Residual network (ResNet) and Inception based CNN model for differentiating the mammograms into the abnormal class and the normal class. Experimental results demonstrated that the deep learners are effective for classifying mammograms and Inception deep learner achieved the best accuracy of 91.49%. [ABSTRACT FROM AUTHOR]

Background and Objective: Early diagnosis of Alzheimerchr('39')s disease (AD) seems necessary due to the high cost of care and treatment, the uncertainty of existing therapies, as well as the worrying future of the patient. This study was conducted to AD diagnosis by MRI images using artificial intelligence methods. Methods: In this research, a computer system for early detection of AD with using machine learning algorithms is presented in the framework of computer-aided process. Conditional random field and Inception deep neural network have been adapted for brain MR images to detect AD. Since hippocampal tissue is one of the first tissues to be affected by AD, hence for the early detection of this disease, the hippocampus was located from other brain tissues firstly and then due to the extent to which this tissue is affected, the diagnosis was made. Conditional random field could accurately extract hippocampal fragments of different shapes in all three brain planes. These components serve as the basis for feature extraction by the deep network. The proposed method was tested on standard ADNI dataset images and its performance was demonstrated. The used Inception network has been trained on the huge ImageNet dataset. One of the important steps is knowledge transfer of the problem under consideration. To facilitate this, data augmentation process was applied according to the shape and structure of the hippocampus. Results: The implemented method in this research, achieved to 98.51% accuracy for two-class classification of "Alzheimer" versus "Normal control" and achieved to 93.41% accuracy for two-class classification of "Mild cognitive impairment" versus "Normal control", which increased by 2.56% and 8.41%, compared with the rival methods, respectively. Conclusion: The achieved results of this study showed that the using of artificial intelligence techniques has highly accurate in diagnosing AD according to MRI images.

Fibroids of the uterus are a common benign tumor affecting women of childbearing age. Uterine fibroids (UF) can be effectively treated with earlier identification and diagnosis. Its automated diagnosis from medical images is an area where deep learning (DL)-based algorithms have demonstrated promising results. In this research, we evaluated state-of-the-art DL architectures VGG16, ResNet50, InceptionV3, and our proposed innovative dual-path deep convolutional neural network (DPCNN) architecture for UF detection tasks. Using preprocessing methods including scaling, normalization, and data augmentation, an ultrasound image dataset from Kaggle is prepared for use. After the images are used to train and validate the DL models, the model performance is evaluated using different measures. When compared to existing DL models, our suggested DPCNN architecture achieved the highest accuracy of 99.8 percent. Findings show that pre-trained deep-learning model performance for UF diagnosis from medical images may significantly improve with the application of fine-tuning strategies. In particular, the InceptionV3 model achieved 90% accuracy, with the ResNet50 model achieving 89% accuracy. It should be noted that the VGG16 model was found to have a lower accuracy level of 85%. Our findings show that DL-based methods can be effectively utilized to facilitate automated UF detection from medical images. Further research in this area holds great potential and could lead to the creation of cutting-edge computer-aided diagnosis systems. To further advance the state-of-the-art in medical imaging analysis, the DL community is invited to investigate these lines of research. Although our proposed innovative DPCNN architecture performed best, fine-tuned versions of pre-trained models like InceptionV3 and ResNet50 also delivered strong results. This work lays the foundation for future studies and has the potential to enhance the precision and suitability with which UF is detected.

Wheat disease detection is crucial for disease diagnosis, pesticide application optimization, disease control, and wheat yield and quality improvement. However, the detection of wheat diseases is difficult due to their various types. Detecting wheat diseases in complex fields is also challenging. Traditional models are difficult to apply to mobile devices because they have large parameters, and high computation and resource requirements. To address these issues, this paper combines the residual module and the inception module to construct a lightweight multiscale CNN model, which introduces the CBAM and ECA modules into the residual block, enhances the model’s attention to diseases, and reduces the influence of complex backgrounds on disease recognition. The proposed method has an accuracy rate of 98.7% on the test dataset, which is higher than classic convolutional neural networks such as AlexNet, VGG16, and InceptionresnetV2 and lightweight models such as MobileNetV3 and EfficientNetb0. The proposed model has superior performance and can be applied to mobile terminals to quickly identify wheat diseases.

Capturing feature information effectively is of great importance in the field of computer vision. With the development of convolutional neural networks, concepts like residual connection and multiple scales promote continual performance gains in diverse deep learning vision tasks. In this paper, novel residual feature-reutilization inception and split-residual feature-reutilization inception are proposed to improve performance on various vision tasks. It consists of four parallel branches, each with convolutional kernels of different sizes. These branches are interconnected by hierarchically organized channels, similar to residual connections, facilitating information exchange and rich dimensional variations at different levels. This structure enables the acquisition of features with varying granularity and effectively broadens the span of the receptive field in each network layer. Moreover, according to the network structure designed above, split-residual feature-reutilization inceptions can adjust the split ratio of the input information, thereby reducing the number of parameters and guaranteeing the model performance. Specifically, in image classification experiments based on popular vision datasets, such as CIFAR10 (97.94%), CIFAR100 (85.91%), Tiny Imagenet (70.54%) and ImageNet (80.83%), we obtain state-of-the-art results compared with other modern models under the premise that the models' sizes are approximate and no additional data is used. • A novel multi-scale convolutional neural network architecture residual feature-reutilization inception and split-residual feature-reutilization inception are proposed to fully utilize features from different scales and enlarge the receptive fields. • Split-residual feature-reutilization inception reduces parameter amount and Flops with acceptable accuracy loss by dividing features into multiple groups with ratios referring to the setting of channel number of residual feature-reutilization inception. [ABSTRACT FROM AUTHOR]

Prestressed Concrete Cylinder Pipe (PCCP) is extensively employed in long-distance water transmission projects, owing to its notable pressure-bearing capacity and durability. Nonetheless, the operation under overpressure conditions poses a risk of structural damage and functional failure to PCCP during its service life. Consequently, accurate prediction of the deformation characteristics of overpressure operation with the structural deformation data under standard water pressure is of great importance for ensuring the safety and reliability of the project. However, it is a challenge to obtain spatiotemporal features for traditional algorithms. As a consequence, a novel deep learning model, i.e. Inception-ResNet-LSTM, is put forward to forecast PCCP deformation. It integrates the Inception, ResNet, and Long Short-Term Memory (LSTM). Based on the prototype experiments, the monitoring was undertaken on the deformations of the inner and outer concrete layers, the steel cylinders, and the prestressed steel wire under internal water pressures of 0 MPa, 0.2 MPa, 0.4 MPa, 0.6 MPa, 0.8 MPa, 1.0 MPa, and 1.15 MPa. In the case study, the deformation data of different structures are input into Inception-ResNet-LSTM to establish a model. For all the structures, the R 2 (coefficient of determination) of the Inception-ResNet-LSTM model exceeds 0.97. Based on the results, it is demonstrated that the model has captured the spatiotemporal features from the monitoring data. In comparison to models based on DL and ML algorithms, Inception-ResNet-LSTM demonstrates superior performance, exhibiting heightened accuracy and enhanced generalization capabilities. It provides a new method for PCCP deformation estimation under overpressure. • The prototype experiments of PCCP for deformation were conducted. • Distributed optical fibers were used to monitor data in the prototype test. • The Inception-ResNet-LSTM model for PCCP deformation is proposed. • Compare the model with DL and ML-based models to show the performance. [ABSTRACT FROM AUTHOR]

• A computer-aided detection system is proposed for mass detection in 3D ABUS images. • The presented system includes two cascaded improved 3D UNets. • The 3D UNets are improved with two modified Inception modules in the encoder path. • Moreover, attention units are placed on the skip connections in the second network. • Results of the proposed system are better than the competing networks. Early detection of breast cancer has a significant effect on reducing its mortality rate. For this purpose, automated three-dimensional breast ultrasound (3-D ABUS) has been recently used alongside mammography. The 3-D volume produced in this imaging system includes many slices. The radiologist must review all the slices to find the mass, a time-consuming task with a high probability of mistakes. Therefore, many computer-aided detection (CADe) systems have been developed to assist radiologists in this task. In this paper, we propose a novel CADe system for mass detection in 3-D ABUS images. The proposed system includes two cascaded convolutional neural networks. The goal of the first network is to achieve the highest possible sensitivity, and the second network's goal is to reduce false positives while maintaining high sensitivity. In both networks, an improved version of 3-D U-Net architecture is utilized in which two types of modified Inception modules are used in the encoder section. In the second network, new attention units are also added to the skip connections that receive the results of the first network as saliency maps. The system was evaluated on a dataset containing 60 3-D ABUS volumes from 43 patients and 55 masses. A sensitivity of 91.48% and a mean false positive of 8.85 per patient were achieved. The suggested mass detection system is fully automatic without any user interaction. The results indicate that the sensitivity and the mean FP per patient of the CADe system outperform competing techniques. [ABSTRACT FROM AUTHOR]

In recent days, the number of road accidents are greatly increased in world wide. The people around the world died due to the rapid road crush. The driver behavior must be detected to neglect the occurrence of road accidents. In this work, the three stage deep learning-based techniques are developed to detect the distracted driver action. The developed model is implemented in three phase to detect the distracted behavior of the driver during the driving process. In the first stage, the images are gathered from standard public resources regarding various drivers while driving. Next, the deep features retrieved using VGG16 and Inception and Xception networks. Then, the weighted fusedfusion is done by gathered features, where the weight parameter is optimized through the Adaptive Gradient-based Optimizer (AGBO) algorithm. Further, the driver behavior classification is done through a Deep Belief Network (DBN) with Radial Basis Functions Neural Networks (RBF), where the hyperparameters of DBN and RBF are optimized through the same AGBO algorithm. Finally, the first set of classified outcomes is attained from the images. In the second stage, the gathered images are fed to the feature extraction stage, where the texture features and motion features are gathered from the images. These collected features are fed to the same optimized DBN with RBF and the second set of classified outcomes with human behaviors has been attained.In the third stage, the gathered image fralgoriom the dataset is fed to the PCA for extracting the features. The gathered features are given to the same optimized DBN with the RBF classifier to get the classified outcome. Later, the classified outcomes from the three stages are fed to the ranking procedure for detecting the distracted driver action. When considering the evaluation indices, the suggested distracted driver prediction model attains 97.67% of accuracy, 97.68% of sensitivity, 97.67% of specificity, and 97.55% of precision value. Thus, the performance of the developed model in the distracted driver prediction is enhanced than the existing approaches. [ABSTRACT FROM AUTHOR]

Thermal imaging is increasingly used in poultry, swine, and dairy animal husbandry to detect disease and distress. In intensive pig production systems, early detection of health and welfare issues is crucial for timely intervention. Using thermal imaging for pig treatment classification can improve animal welfare and promote sustainable pig production. In this paper, we present a depthwise separable inception subnetwork (DISubNet), a lightweight model for classifying four pig treatments. Based on the modified model architecture, we propose two DISubNet versions: DISubNetV1 and DISubNetV2. Our proposed models are compared to other deep learning models commonly employed for image classification. The thermal dataset captured by a forward-looking infrared (FLIR) camera is used to train these models. The experimental results demonstrate that the proposed models for thermal images of various pig treatments outperform other models. In addition, both proposed models achieve approximately 99.96–99.98% classification accuracy with fewer parameters.

Automatic road extraction from remote sensing images has an important impact on road maintenance and land management. While significant deep-learning-based approaches have been developed in recent years, achieving a suitable trade-off between extraction accuracy, inference speed and model size remains a fundamental and challenging issue for real-time road extraction applications, especially for rural roads. For this purpose, we developed a lightweight dynamic addition network (LDANet) to exploit rural road extraction. Specifically, considering the narrow, complex and diverse nature of rural roads, we introduce an improved Asymmetric Convolution Block (ACB)-based Inception structure to extend the low-level features in the feature extraction layer. In the deep feature association module, the depth-wise separable convolution (DSC) is introduced to reduce the computational complexity of the model, and an adaptation-weighted overlay is designed to capture the salient features. Moreover, we utilize a dynamic weighted combined loss, which can better solve the sample imbalance and boosts segmentation accuracy. In addition, we constructed a typical remote sensing dataset of rural roads based on the Deep Globe Land Cover Classification Challenge dataset. Our experiments demonstrate that LDANet performs well in road extraction with fewer model parameters (

In deeper layers, ResNet heavily depends on skip connections and Relu. Although skip connections have demonstrated their usefulness in networks, a major issue arises when the dimensions between layers are not consistent. In such cases, it is necessary to use techniques such as zero-padding or projection to match the dimensions between layers. These adjustments increase the complexity of the network architecture, resulting in an increase in parameter number and a rise in computational costs. Another problem is the vanishing gradient caused by utilizing Relu. In our model, after making appropriate adjustments to the inception blocks, we replace the deeper layers of ResNet with modified inception blocks and Relu with our non-monotonic activation function (NMAF). To reduce parameter number, we use symmetric factorization and 1×1 convolutions. Utilizing these two techniques contributed to reducing the parameter number by around 6 M parameters, which has helped reduce the run time by 30 s/epoch. Unlike Relu, NMAF addresses the deactivation problem of the non-positive number by activating the negative values and outputting small negative numbers instead of zero in Relu, which helped in enhancing the convergence speed and increasing the accuracy by 5%, 15%, and 5% for the non-noisy datasets, and 5%, 6%, 21% for non-noisy datasets.

In recent times, Artificial Intelligence (AI) attracted many researchers in medical image analysis. Analyzing the vast medical data through traditional approaches is a bit tedious and time-consuming in designing feature descriptors. Therefore, we present a Generative Squeeze Net-UNet framework for image retrieval in this paper. Proposed framework consists of an image reconstruction module using an encoder with series of six encoder blocks taken from Squeeze Net and a sequential decoder from UNet with transpose fire modules. Also, the image matching module is followed by retrieval of similar images from database. Encoder module of the proposed framework consists of encoder blocks with Squeeze Net architecture that encode the features from a given image with different kernel size and are forwarded to decoder module. Subsequently, reconstruction of image takes place. Extracted features form the basis for reconstruction of input image. Encoded latent features are used for image matching and retrieval of similar images from a database. Performance of the Generative Squeeze Net-UNet framework is analyzed using standard image databases like VIA/ELCAP)CT and OASIS. Proposed method is compared with the state-of-the-art deep learning as well as local descriptor approaches. Average precision and recall are calculated against two datasets. Metrics suchas d1 and Euclidian distances are used in the process. Experimental results show that the proposed framework outperforms the existing works. [ABSTRACT FROM AUTHOR]

Manual segmentation of brain tumor is not only a tedious task that may bring human mistakes. An automatic segmentation gives results faster, and it extends the survival rate with an earlier treatment plan. So, an automatic brain tumor segmentation model, modified inception module based U-Net (IMU-Net) proposed. It takes Magnetic resonance (MR) images from the BRATS 2017 training dataset with four modalities (FLAIR, T1, T1ce, and T2). The concatenation of two series 3×3 kernels, one 5×5, and one 1×1 convolution kernels are utilized to extract the whole tumor (WT), core tumor (CT), and enhance tumor (ET). The modified inception module (IM) collects all the relevant features and provides better segmentation results. The proposed deep learning model contains 40 convolution layers and utilizes intensity normalization and data augmentation operation for further improvement. It achieved the mean dice similarity coefficient (DSC) of 0.90, 0.77, 0.74, and the mean Intersection over Union (IOU) of 0.79, 0.70, 0.70 for WT, CT, and ET during the evaluation. [ABSTRACT FROM AUTHOR]

With the rapid growth of the number of flights, the traditional radar system has been unable to meet the needs of flight supervision. At the same time, it also puts forward higher requirements for air traffic management (ATM). Automatic dependent surveillance-broadcast (ADS-B) is a promising technology in the next generation of air traffic control (ATC). However, the openness of ADS-B system brings the opportunities for terrorists to tamper with data. In this paper, we propose a novel aircraft coordinate prediction hybrid model based on deep learning. The proposed model combines inception modules and long short-term memory (LSTM) modules. Inception modules are used to extract the spatial features of dataset, and LSTM modules are used to extract the temporal features of dataset. In addition, we use the ADS-B signal strength instead of its specific information to obtain aircraft coordinates. Signal strength is not easily tampered with, but it carries limited information. Therefore, this scheme sacrifices a certain precision for reliability. Inception modules and LSTM modules are combined in different ways to perform experiments on the real-world ADS-B datasets from OpenSky network. The experimental results show that the proposed 2-Inception-LSTM is the local optimal model. The prediction error is within 10 km. It can be suitable for situations where the positioning accuracy of aircraft coordinates is not pursued, but the positioning reliability must be guaranteed. [ABSTRACT FROM AUTHOR]

Neural networks traditionally produce a scalar value for an activated neuron. Capsules, on the other hand, produce a vector of values, which has been shown to correspond to a single, composite feature wherein the values of the components of the vectors indicate properties of the feature such as transformation or contrast. We present a new way of parameterizing and training capsules that we refer to as homogeneous vector capsules (HVCs). We demonstrate, experimentally, that altering a convolutional neural network (CNN) to use HVCs can achieve superior classification accuracy without increasing the number of parameters or operations in its architecture as compared to a CNN using a single final fully connected layer. Additionally, the introduction of HVCs enables the use of adaptive gradient descent, reducing the dependence a model’s achievable accuracy has on the finely tuned hyperparameters of a non-adaptive optimizer. We demonstrate our method and results using two neural network architectures. For the CNN architecture referred to as Inception v3, replacing the fully connected layers with HVCs increased the test accuracy by an average of 1.32% across all experiments conducted. For a simple monolithic CNN, we show HVCs improve test accuracy by an average of 19.16%.

Traffic sign recognition with high accuracy and real-time is an important part of the intelligent transportation system. In this article, based on large-scale traffic signs and the inherent conflict between location regression and classification of traffic signs, we propose a novel and flexible two-stage approach. It combines a lightweight superclass detector with a refinement classifier. The main contributions lie in three aspects: (1) We use locations and sizes of signs as prior knowledge to establish a probability distribution model. It can significantly decrease the search range of signs and improve the processing speed, as well as reducing false detection. (2) We propose a high-performance lightweight superclass detector. We introduce the Inception and Channel Attention, by generating multi-scale receptive fields and adaptively adjusting channel features. It alleviates the large scale variance challenge of objects and the interference of background information. Meanwhile, we present a merging Batch Normalization and multi-scale testing method to further improve detection performance. (3) We propose a refinement classifier based on similarity measure learning for the subclass classification. It increases the precision of discriminating similar subclasses and also improves the extensibility of our approach. Our two-stage approach is simple and effective, whose paradigm is different from others. Experiments on the Tsinghua-Tencent 100K dataset demonstrate the performance of our approach. Compared with the state-of-the-art methods, our method achieves competitive performance (92.16% mAP) with a lightweight detector ($6.49M $). The processing time is $0.150s $ per frame, of which the speed is increased by 3 times compared with existing methods. [ABSTRACT FROM AUTHOR]

Agriculture is one of the primary occupations in many countries. Tomatoes are grown by many farmers in countries where the water resource is available in abundance. Improper methods of cultivation and failure to identify the diseases when it is in the nascent stage results in the reduction of crop yield thus affecting the outcome of cultivation. This paper proposes a novel method of early identification of diseases in tomato plants by making use of convolutional neural networks (CNN) and image processing. Dataset from an open repository was considered for training and testing and the algorithm was capable of identifying nine different varieties of diseases that affect the tomato plant at its early stages. The images of tomato leaves were fed for identification through processing and classification. An optimum model was developed by analyzing various architectures of CNN including the VGG, ResNet, Inception, Xception, MobileNet and DenseNet. The performance of each of these architectures was compared and various metrics like the accuracy, loss, precision, recall and area under the curve (AUC) were analyzed. [ABSTRACT FROM AUTHOR]

To secure cloud computing and outsourced data while meeting the requirements of automation, many intrusion detection schemes based on deep learning are proposed. Though the detection rate of many network intrusion detection solutions can be quite high nowadays, their identification accuracy on imbalanced abnormal network traffic still remains low. Therefore, this paper proposes a ResNet &Inception-based convolutional neural network (RICNN) model to abnormal traf- fic classification. RICNN can learn more traffic features through the Inception unit, and the degradation problem of the network is eliminated through the direct mapping unit of ResNet, thus the improvement of the model's generalization ability can be achievable. In addition, to simplify the network, an improved version of RICNN, which makes it possible to reduce the number of parameters that need to be learnt without degrading identification accuracy, is also proposed in this paper. The experimental results on the dataset CICIDS2017 show that RICNN not only achieves an overall accuracy of 99.386% but also has a high detection rate across different categories, especially for small samples. The comparison experiments show that the recognition rate of RICNN outperforms a variety of CNN models and RNN models, and the best detection accuracy can be achieved. [ABSTRACT FROM AUTHOR]

The images classification is one of the most common applications of deep learning. Images of dogs and cats are mostly used as examples for image classification models, as they are relatively easy for the human eyes to recognize. However, classifying the breed of a dog or a cat has its own complexity. In this paper, a fine-tuned pre-trained model of a Keras' application was built with a new dataset of dogs and cats to predict the breed of identified dogs or cats. Keras applications are deep learning models, which have been previously trained with general image datasets from ImageNet. In this paper, the ResNet-152 v2, Inception-ResNet v2, and Xception models, adopted from Keras application, are retrained to predict the breed among the 21 classes of dogs and cats. Our results indicate that the Xception model has produced the highest prediction accuracy. The training accuracy is 99.49%, the validation accuracy is 99.21%, and the testing accuracy is 91.24%. Besides, the training time is about 14 hours and the predicting time is about 18.41 seconds. [ABSTRACT FROM AUTHOR]

Accurate detection and pixel‐wise classification of brain tumors in Magnetic Resonance Imaging (MRI) scans are vital for their diagnosis, prognosis study and treatment planning. Manual segmentation of tumors from MRI is highly subjective and tedious. With recent advances in deep learning, automatic brain tumor segmentation is an emerging research direction in the medical imaging domain. We present a study to improve the automatic segmentation process by introducing size variability in the Convolutional Neural Network (CNN). For pixel‐wise classification of tumorous slices convolutional neural network‐based encoder‐decoder UNET model is referred. A multi‐inception‐UNET model is proposed to improve scalability of the UNET model. Extensive experiments have been performed using the Brain Tumor Segmentation Challenge (BRATS) datasets to establish the validity of our proposed model. Experimental results show that our proposed method achieved the best results on BraTS 2015, 2017 and 2019 datasets for complete tumor, core tumor and enhancing tumor regions respectively. [ABSTRACT FROM AUTHOR]

Brain tumor is a severe cancer disease caused by uncontrollable and abnormal partitioning of cells. Recent progress in the field of deep learning has helped the health industry in Medical Imaging for Medical Diagnostic of many diseases. For Visual learning and Image Recognition, task CNN is the most prevalent and commonly used machine learning algorithm. Similarly, in our paper, we introduce the convolutional neural network (CNN) approach along with Data Augmentation and Image Processing to categorize brain MRI scan images into cancerous and non-cancerous. Using the transfer learning approach we compared the performance of our scratched CNN model with pre-trained VGG-16, ResNet-50, and Inception-v3 models. As the experiment is tested on a very small dataset but the experimental result shows that our model accuracy result is very effective and have very low complexity rate by achieving 100% accuracy, while VGG-16 achieved 96%, ResNet-50 achieved 89% and Inception-V3 achieved 75% accuracy. Our model requires very less computational power and has much better accuracy results as compared to other pre-trained models.

With the continuous development of the earth observation technology, the spatial resolution of remote sensing images is also continuously improved. As one of the key problems in remote sensing images interpretation, the classification of high-resolution remote sensing images has been widely concerned by scholars at home and abroad. With the improvement of science and technology, deep learning has provided new ideas for the development of image classification, but it has not been widely used in remote sensing images processing. In the background of remote sensing huge data, the remote sensing images classification based on deep learning proposed in the study has more research significance and application value. The study proposes a high-resolution remote sensing images classification method based on an improved convolutional neural network. The traditional convolutional neural network framework is optimized and the initial structure is added. The actual classification results of radial basis functions and support vector machine are compared horizontally. The classification results of hyperspectral images were presented that the improved method can perform better in overall accuracy and Kappa coefficient. The commission errors of support vector machine classification method are more than 6 times of that of the improved convolutional neural network classification method and the overall accuracy of the improved convolutional neural network classification method has reached 97% above.

In this paper, a new CNN architecture is introduced for classification of six different hand gestures using surface electromyography (EMG) data collected from the forearm. At first, two different deep neural networks produced based on Slow Fusion and Inception models separately. Then, the average of accuracy values and standard deviations were calculated for each type of network. The average accuracy was 80.88% and standard deviation was 0.030 for the Slow Fusion based network. For the Inception based network, average accuracy was 82.64% and standard deviation was 0.028. In addition to these two networks, a new CNN architecture is introduced using Slow fusion and Inception models in combination. The architecture has two parallel Inception modules in parallel. Each parallel module is fed by the half of the 3D feature map. The proposed model slowly fuses the information of the parallel modules throughout the network as in Slow-Fusion architecture. The average accuracy achieved with this model was 83.97% and the standard deviation was 0.027. Despite the small data set, the accuracy had increased with the proposed hybrid model. The smaller standard deviation indicates that it is less affected by variations in the training dataset. Our experimental results show that the proposed method gives the best results among the Slow Fusion based and Inception based models.

Accurate damage location diagnosis of frame structures is of great significance to the judgment of damage degree and subsequent maintenance of frame structures. However, the similarity characteristics of vibration data at different damage locations and noise interference bring great challenges. In order to overcome the above problems and realize accurate damage location diagnosis of the frame structure, the existing convolutional neural network with training interference (TICNN) is improved in this paper, and a high-precision neural network model named convolutional neural network based on Inception (BICNN) for fault diagnosis with strong anti-noise ability is proposed by adding the Inception module to TICNN. In order to effectively avoid the overall misjudgment problem caused by using single sensor data for damage location diagnosis, an integrated damage location diagnosis method is proposed. Taking the four-story steel frame model of the University of British Columbia as the research object, the method proposed in this paper is tested and compared with other methods. The experimental results show that the diagnosis accuracy of the proposed method is 97.38%, which is higher than other methods; at the same time, it has greater advantages in noise resistance. Therefore, the method proposed in this paper not only has high accuracy, but also has strong anti-noise ability, which can solve the problem of accurate damage location diagnosis of complex frame structures under a strong noise environment.

Accurate brain tumor segmentation is crucial for clinical assessment, follow-up, and subsequent treatment of gliomas. While convolutional neural networks (CNN) have become state of the art in this task, most proposed models either use 2D architectures ignoring 3D contextual information or 3D models requiring large memory capacity and extensive learning databases. In this study, an ensemble of two kinds of U-Net-like models based on both 3D and 2.5D convolutions is proposed to segment multimodal magnetic resonance images (MRI). The 3D model uses concatenated data in a modified U-Net architecture. In contrast, the 2.5D model is based on a multi-input strategy to extract low-level features from each modality independently and on a new 2.5D Multi-View Inception block that aims to merge features from different views of a 3D image aggregating multi-scale features. The Asymmetric Ensemble of Asymmetric U-Net (AE AU-Net) based on both is designed to find a balance between increasing multi-scale and 3D contextual information extraction and keeping memory consumption low. Experiments on 2019 dataset show that our model improves enhancing tumor sub-region segmentation. Overall, performance is comparable with state-of-the-art results, although with less learning data or memory requirements. In addition, we provide voxel-wise and structure-wise uncertainties of the segmentation results, and we have established qualitative and quantitative relationships between uncertainty and prediction errors. Dice similarity coefficient for the whole tumor, tumor core, and tumor enhancing regions on BraTS 2019 validation dataset were 0.902, 0.815, and 0.773. We also applied our method in BraTS 2018 with corresponding Dice score values of 0.908, 0.838, and 0.800. [ABSTRACT FROM AUTHOR]

Copyright of Journal of Computer Engineering & Applications is the property of Beijing Journal of Computer Engineering & Applications Journal Co Ltd. and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Accurate brain tumor segmentation is crucial for clinical assessment, follow-up, and subsequent treatment of gliomas. While convolutional neural networks (CNN) have become state of the art in this task, most proposed models either use 2D architectures ignoring 3D contextual information or 3D models requiring large memory capacity and extensive learning databases. In this study, an ensemble of two kinds of U-Net-like models based on both 3D and 2.5D convolutions is proposed to segment multimodal magnetic resonance images (MRI). The 3D model uses concatenated data in a modified U-Net architecture. In contrast, the 2.5D model is based on a multi-input strategy to extract low-level features from each modality independently and on a new 2.5D Multi-View Inception block that aims to merge features from different views of a 3D image aggregating multi-scale features. The Asymmetric Ensemble of Asymmetric U-Net (AE AU-Net) based on both is designed to find a balance between increasing multi-scale and 3D contextual information extraction and keeping memory consumption low. Experiments on 2019 dataset show that our model improves enhancing tumor sub-region segmentation. Overall, performance is comparable with state-of-the-art results, although with less learning data or memory requirements. In addition, we provide voxel-wise and structure-wise uncertainties of the segmentation results, and we have established qualitative and quantitative relationships between uncertainty and prediction errors. Dice similarity coefficient for the whole tumor, tumor core, and tumor enhancing regions on BraTS 2019 validation dataset were 0.902, 0.815, and 0.773. We also applied our method in BraTS 2018 with corresponding Dice score values of 0.908, 0.838, and 0.800.

Building segmentation is crucial for applications extending from map production to urban planning. Nowadays, it is still a challenge due to CNNs’ inability to model global context and Transformers’ high memory need. In this study, 10 CNN and Transformer models were generated, and comparisons were realized. Alongside our proposed Residual-Inception U-Net (RIU-Net), U-Net, Residual U-Net, and Attention Residual U-Net, four CNN architectures (Inception, Inception-ResNet, Xception, and MobileNet) were implemented as encoders to U-Net-based models. Lastly, two Transformer-based approaches (Trans U-Net and Swin U-Net) were also used. Massachusetts Buildings Dataset and Inria Aerial Image Labeling Dataset were used for training and evaluation. On Inria dataset, RIU-Net achieved the highest IoU score, F1 score, and test accuracy, with 0.6736, 0.7868, and 92.23%, respectively. On Massachusetts Small dataset, Attention Residual U-Net achieved the highest IoU and F1 scores, with 0.6218 and 0.7606, and Trans U-Net reached the highest test accuracy, with 94.26%. On Massachusetts Large dataset, Residual U-Net accomplished the highest IoU and F1 scores, with 0.6165 and 0.7565, and Attention Residual U-Net attained the highest test accuracy, with 93.81%. The results showed that RIU-Net was significantly successful on Inria dataset. On Massachusetts datasets, Residual U-Net, Attention Residual U-Net, and Trans U-Net provided successful results.

Fault diagnosis of bearing with small data samples is always a research hotspot in the field of bearing fault diagnosis. To solve the problem, a convolutional block attention module (CBAM)-based attentional feature fusion with an inception module based on a capsule network (Capsnet) is proposed in the paper. Firstly, the original vibration signal is decomposed into multiple intrinsic mode function (IMF) sub-signals by the ensemble empirical mode decomposition (EEMD), and then the original vibration signal and the corresponding former four order IMF sub-signals are input into the inception modules to extract the features. Secondly, these features are concatenated and optimized by the CBAM. Finally, the selected sensitive features are fed into the Capsnet to diagnose the faults. Through the multifaceted experiment analysis on fault diagnosis of bearing with small data samples, the diagnosis results demonstrate that the proposed attentional feature fusion with inception based on Capsnet not only diagnoses the fault of bearing with small data samples, but also is superior to other feature fusion methods, such as feature fusion with inception based on Capsnet and attentional feature fusion with inception based on CNN, etc., and other single diagnosis models such as Capsnet with CBAM and inception, and CNN with CBAM and inception.

This paper is an experimental study of the electrical conduction of mineral-oil/solid interfaces in negative polarity. This study is performed in a point-plane configuration immersed in mineral oil submitted to impulse voltages. The interface is assembled with an inclined solid in contact with the point electrode tip. Charge recordings and conduction currents of different combinations of mineral-oil/solid interfaces are reported. Different polymers and oil-impregnated papers are used as the solid materials. It is found that for voltages lower than the inception voltage, the injection of charge belongs to a weak injection regime. The spatial limitation with solids of similar permittivity as the mineral oil decreases the charge injection. The charge injection at mineral-oil/solid interfaces follow two conduction current type characteristics. Experiments with mineral-oil/PTFE interface shows the existence of a transition between both conduction current types if the applied voltage is increased. [ABSTRACT FROM AUTHOR]

Convolutional neural network (CNN) models have been used extensively in many image recognition tasks for their state-of-the-art performance in recent years. Researchers inspired by this success frequently prefer CNNs in the agricultural field, especially for disease detection and classification. Many CNN models have been proposed for plant leaf diseases and impressive performance results have been obtained. On the other hand, standard CNNs usually need millions of parameters in the network for computation, but it is difficult to implement them on embedded and mobile devices with limited resources. Therefore, it is important to obtain lighter models by decreasing the number of parameters in addition to the high performance of the models. In this paper, a new hybrid CNN approach based on Inception architecture and depthwise separable convolutions is proposed to reduce the number of parameters and computational cost for plant leaf disease detection and classification. Although the number of parameters is significantly reduced, the results show that the proposed approach has high accuracy performance. The proposed hybrid model has been trained and tested with k-fold cross-validation using a dataset of 50,136 images containing 30 classes from 14 different leaves, including healthy and diseased ones. The new model has achieved the best accuracy of 99.27% and an average accuracy of 99%, and provides about a 75% reduction in the number of parameters compared to the standard CNN. [ABSTRACT FROM AUTHOR]

The variable working conditions and frequent turns make the aircraft actuator system prone to failure, seriously threatening flight safety. The identification of the airplane actuator system is critical for flight decisions and safety. Most fault diagnosis methods of actuators only focus on the actuators themselves, ignoring the disturbance caused by the fault of the actuator position sensor, which may easily lead to wrong decisions. In order to distinguish the actuator fault from its position sensor fault and identify the fault type accurately, an offline diagnosis method of convolutional neural network (CNN) with novel topology for processing time series is proposed. A new shift layer is added after the input layer, which avoids the loss of a large number of features due to the direct connection between the time series and the convolutional layer. A local topological network learning complex pattern with inception module is designed to improve the diagnostic accuracy in different working conditions. The wide residual structure is introduced to expand the convolutional channel, which allows the network features at the bottom level to propagate directly to the top level to prevent network degradation. Simulation results show that this method can accurately diagnose the actuator fault and its position sensor, with an average accuracy of 96.8%. Compared with the current mainstream data-driven methods, the precision and recall are increased by 6.3% and 6.7% respectively. [ABSTRACT FROM AUTHOR]

Hyperspectral imaging (HSI) classification has recently become a field of interest in the remote sensing (RS) community. However, such data contain multidimensional dynamic features that make it difficult for precise identification. Also, it covers structurally nonlinear affinity within the gathered spectral bands and the related materials. To systematically facilitate the HSI categorization, we propose a spectral-spatial classification of HSI data using a 3D-2D convolutional neural network and inception network to extract and learn the in-depth spectral-spatial feature vectors. We first applied the principal component analysis (PCA) on the entire HSI image to reduce the original space dimensionality. Second, the exploitation of the spatial hyperspectral input features contiguous information by 2-D CNN. Besides, we used 3-D CNN without relying on any preprocessing to extract deep spectral-spatial fused features efficiently. The learned spectral-spatial characteristics are concatenated and fed to the inception network layer for joint spectral-spatial learning. Furthermore, we learned and achieved the correct classification with a softmax regression classifier. Finally, we evaluated our model performance on different training set sizes of two hyperspectral remote sensing data sets (HSRSI), namely Botswana (BT) and Kennedy Space Center (KSC), and compared the experimental results with deep learning-based and state-of-the-art (SOTA) classification methods. The experiment results show that our model provides competitive classification results with state-of-the-art techniques, demonstrating the considerable potential for HSRSI classification. [ABSTRACT FROM AUTHOR]