Search

Document falsification is among the fastest growing problems all over the world. Disclosure of such document is not always possible due to the conspiracy of attorney bodies; especially legal documents such as bank cheques, contracts, cash memos, and so on. Handwritten document tampering detection due to addition of new word(s) in judicial documents is the prime objective of this research. Minute alteration in writing causes financial loss to a person or to an organization and decreases the global economy. Such intangible assets remain undiscovered owing to lack of proper forensic techniques. Though writing style imitation can be possible, however, the possibility of getting exactly the same pen of the authorized document is quite impossible for an imitator. Hence, the paper introduces a solution to detect forgery in handwritten legal documents by analyzing perceptually similar pen ink. Forgery activity happens either ends of a written document by appending new word(s)/letter(s) with similar type of pen. The work is formulated as a binary classification problem and established with the help of several statistical features and three different classifiers: Multilayer Perceptron(MLP), RBF-SVM, and Random Forest(RF). Besides, the problem has also been implemented through some DCNN approaches to check whether it is possible to reflect the forgery by direct approaches. The efficiency of the proposed method is quite promising for involvement in the examination of forensic documents. [ABSTRACT FROM AUTHOR]

The Sundarban constitute the world's largest continuous mangrove forest comprising of the natural habitat of the Royal Bengal Tiger. There is an urgent need for developing a system that would help in monitoring, identifying, and counting the number of tigers not only ensuring the protection of the tigers but also prioritizing the least human disturbance and intervention to the physically inaccessible areas. In this paper, a system is proposed for the identification and detection of Royal Bengal Tigers using drone image-capturing with an active GPS and thermal detection. In this proposed work, the drone with an active GPS will help in monitoring the location and thermal scan to track the tigers. The captured image of the tiger by the drone will be passed to a Deep Convolutional Neural Network (DCNN) using YOLO and Faster RCNN that will help in the detection of the uncommon features of each tiger hence identifying and counting the number of tigers in a region. [ABSTRACT FROM AUTHOR]

With the continuous development of deep learning, Deep Convolutional Neural Network (DCNN) has attracted wide attention in the industry due to its high accuracy in image classification. Compared with other DCNN hardware deployment platforms, Field Programmable Gate Array (FPGA) has the advantages of being programmable, low power consumption, parallelism, and low cost. However, the enormous amount of calculation of DCNN and the limited logic capacity of FPGA restrict the energy efficiency of theDCNNaccelerator. The traditional sequential sliding window method can improve the throughput of the DCNN accelerator by data multiplexing, but this method’s datamultiplexing rate is low because it repeatedly reads the data between rows. This paper proposes a fast data readout strategy via the circular sliding window data reading method, it can improve the multiplexing rate of data between rows by optimizing the memory access order of input data. In addition, themultiplication bit width of the DCNN accelerator ismuch smaller than that of the Digital Signal Processing (DSP) on the FPGA, whichmeans that there will be a waste of resources if amultiplication uses a single DSP. Amultiplier sharing strategy is proposed, the multiplier of the accelerator is customized so that a singleDSP block can complete multiple groups of 4, 6, and 8-bit signed multiplication in parallel. Finally, based on two strategies of appeal, an FPGA optimized accelerator is proposed.The accelerator is customized byVerilog language and deployed onXilinxVCU118.When the accelerator recognizes the CIRFAR-10 dataset, its energy efficiency is 39.98 GOPS/W, which provides 1.73 × speedup energy efficiency over previous DCNN FPGA accelerators. When the accelerator recognizes the IMAGENET dataset, its energy efficiency is 41.12 GOPS/W, which shows 1.28×−3.14 × energy efficiency compared with others. [ABSTRACT FROM AUTHOR]

One of the unsolved problems in computer vision is recognizing or understanding other people's emotions and feelings. Deep Convolutional Neural Networks (CNNs) have attempted to be effective in addressing emotion recognition issues. The significant level of performance achieved by these classifiers can be attributed to their ability to self-learn a downsampled feature vector that retains abstraction information through filter kernels in convolutional layers. Despite these advancements, challenges persist in capturing subtle and context-dependent emotional nuances, hindering the development of more nuanced emotion recognition systems. Ongoing research explores multimodal approaches, integrating various sensory inputs, to enhance the accuracy and reliability of emotion detection in real-world scenarios. In this paper, we explore the impact of training the initial weights in an unsupervised manner. We study the results of pretraining a Deep CNN as a Convolutional Auto-Multiplexer (CAM) in a greedy layer-wise unsupervised fashion for emotion recognition using facial feature images. When trained with randomly initialized weights, our CNN emotion recognition model achieves a performance rate of 92.16% on the Karolinska Directed Emotional Faces (KDEF) dataset. In contrast, by using this pre-trained model, the performance increases to 93.52%. Pre-training our CNN as a CAM also marginally reduces training time. [ABSTRACT FROM AUTHOR]

With the development of wireless communication technology, the electromagnetic interference (EMI) of artificial radio to weather radar increases significantly, which has a serious impact on the quality of radar data. Most of the research on detecting and suppressing electromagnetic interference was based on the primary product of the radar. This paper researches the effectiveness of deep convolutional neural networks (DCNN) to identify and suppress electromagnetic interference based on the I/Q data output from the front end of a radar receiver. Firstly, this paper selected UNet, ResNet with UNet structure, and DeepLab V3+ for the semantic segmentation of electromagnetic interference and other signals. After semantic segmentation, this paper used the linear interpolation method to suppress EMI. Finally, this paper selected the prediction precision of the model and compared the quality of primary products before and after EMI suppression to evaluate the effectiveness of DCNN. The results showed that all three models could effectively identify the electromagnetic interference and the quality of the data were improved after suppression. It suggests that the use of DCNN on the I/Q data output from the front end of a radar receiver can play a certain effect on the identification of electromagnetic interference. [ABSTRACT FROM AUTHOR]

Recently, there had been a massive group of people, who were being rapidly affected by melanoma. Melanoma is a form of skin cancer that develops on the skin's surface layer. This is primarily caused due to excessive skin exposure to UV radiation and severe sunburns. Thus, the early detection of melanoma can aid us to cure it completely. This paper intends to introduce a new melanoma detection framework with four main phases viz. segmentation, feature extraction, optimal feature selection, as well as detection. Initially, the segmentation process takes place to the input skin image via Fuzzy C-Means Clustering (FCM) approach. From the segmented image ( Im seg ) , some of the features such as Gray Level Run Length Matrix (GLRM), Local Vector Pattern (LVP), Local Binary Pattern (LBP), Local Directional Pattern (LDP) and Local Tetra Pattern (LTrP) are extracted. As the extracted features (F) suffered from the issue of "curse of dimensionality", this paper utilizes optimization to select optimal features, which makes the detection more precise. As a novelty, a new hybrid algorithm Particle-Assisted Moth Search Algorithm (PA-MSA) is introduced that hybridizes the concept of Moth Search Algorithm (MSA) and Particle Swarm Optimization (PSO), respectively. For the classification process, the optimally chosen features (F opt ) are fed as input, where Deep Convolution Neural Network (DCNN) is used. Finally, a performance-based comparative analysis is conducted among the proposed PA-MSA as well as the existing models with respect to various measures. [ABSTRACT FROM AUTHOR]

Retinopathy is an eye disease caused by diabetes, and early detection and treatment can potentially reduce the risk of blindness in diabetic retinopathy sufferers. Using retinal Fundus images, diabetic retinopathy can be diagnosed, recognized, and treated. In the current state of the art, sensitivity and specificity are lacking. However, there are still a number of problems to be solved in state-of-the-art techniques like performance, accuracy, and being able to identify DR disease effectively with greater accuracy. In this paper, we have developed a new approach based on a combination of image processing and artificial intelligence that will meet the performance criteria for the detection of disease-causing diabetes retinopathy in Fundus images. Automatic detection of diabetic retinopathy has been proposed and has been carried out in several stages. The analysis was carried out in MATLAB using software-based simulation, and the results were then compared with those of expert ophthalmologists to verify their accuracy. Different types of diabetic retinopathy are represented in the experimental evaluation, including exudates, micro-aneurysms, and retinal hemorrhages. The detection accuracies shown by the experiments are greater than 98.80 percent. [ABSTRACT FROM AUTHOR]

Traditional models for pose estimation in video surveillance are based on graph structures, in this paper, we propose a method that breaks the limitation of template matching within a range of pose changes to obtain robust results. We implement our swimmer pose estimation method based on deep learning. We take use of High-Resolution Net (HRNet) to extract and fuse visual features of visual object and complete the object detection using the key points of human joint. The proposed model could be applied to all kinds of swimming styles throughout appropriate training. Compared with the methods that require multimodel combinations and training, the proposed method directly achieves the end-to-end prediction, which is easily to be implemented and deployed. In addition, a cross-fusion module is added between parallel networks, which assists the network to make use of the characteristics of multiple resolutions. The proposed network has achieved ideal results in the pose estimation of swimmers by comparing HRNet-W32 and HRNet-W48. In addition, we propose an annotated key point dataset of swimmers which was created from the view of underwater swimmers. Compared with side view, the torso of swimmers collected by the underwater view is much suitable for a broad spectrum of machine vision tasks. [ABSTRACT FROM AUTHOR]

Segmentation of skin lesions is a significant and demanding task in dermoscopy images. This paper proposes a new skin cancer recognition scheme, with: "Pre-processing, Segmentation, Feature extraction, Optimal Feature Selection and Classification". Here, pre-processing is done with certain processes. The pre-processed images are segmented via the "Otsu Thresholding model". The third phase is feature extraction, where Deviation Relevance-based "Local Binary Pattern (DRLBP), Gray-Level Co-Occurrence Matrix (GLCM) features and Gray Level Run-Length Matrix (GLRM) features" are extracted. From these extracted features, the optimal features are chosen via Particle Updated WOA (PU-WOA) model. Subsequently, classification occurs via Optimized DCNN and NN to classify the skin lesion. To make the classification more precise, the DCNN is optimized by the introduced algorithm. The result has shown a higher accuracy of 0.998737, when compared with other extant models like IPSO, IWOA, PSO+CNN, WOA+CNN and CNN schemes. [ABSTRACT FROM AUTHOR]

Image super resolution (SR) is a critical category within the field of image processing techniques that aims to improve the resolution of both images and videos. Its significance is especially pronounced in image processing and computer vision domain. In recent years, deep learning approaches for image super resolution have consistently achieved remarkable advancements. This paper presents a new approach for enhancing the quality of images. More precisely, we have contributed in the single image super-resolution field by exploiting beta wavelet in deep convolutional neural network (DCNN), which we called deep convolutional neural network for image super resolution boosting using beta wavelet (ISR2BW). First, the low-resolution image undergoes a division into four sub-bands, consisting of one approximation and three distinct detail components. Second, various analytic beta wavelet (ABW) filters are computed to forecast and restore the missing details within the approximation sub-band. Then, this output is used as input to the DCNN in order to improve SR wavelet coefficients. The predicted approximation and three sub-bands are subsequently subjected to the inverse 2D discrete wavelet transformation to generate the high image super resolution. The proposed method is assessed through extensive experiments by examining the structural similarity index for measure, the peak signal to noise ratio, and the mean squared error indicators. The experimental results highlight the ISR2BW method's effectiveness and efficiency when compared to other state-of-the-art methods on commonly used benchmarking datasets. [ABSTRACT FROM AUTHOR]

The carrier-free UWB sensor features high distance resolution and high interference immunity. It is not easily affected by weather and lighting conditions, and its received echoes contain detailed structural information of the target. This paper proposes a small sample terrain recognition framework based on the carrier-free UWB sensor. The time-frequency feature maps of terrain echo signals are used for classification. However, insufficient samples make the classifier prone to overfitting, so we propose an Improved Auxiliary Classifier Generative Adversarial Network (IACGAN) for data enhancement in this paper. Firstly, attention mechanism and multi-scale convolution are added to the network structure of ACGAN to improve the feature extraction capability of time-feature images of echo signals. Secondly, the discriminator’s true/false judgment criterion changes from Jensen-Shannon divergence to Wasserstein distance with gradient penalty, improving training stability. Finally, label classification of the generated samples by the discriminator is eliminated, which further enhances the quality of the generated images. Experiments show that the IACGAN improves the quality of generated images with IS and FID as the generation quality evaluation criteria. Furthermore, k-fold cross-validation shows that data augmentation by IACGAN improves the recognition rate of the CNN classifier. Finally, the experiment also found that directly using the discriminator in the trained IACGAN as the classifier can achieve more than 97% accuracy. That does not require additional training of the classifier on the expanded training set, which is an efficient and low-cost alternative. [ABSTRACT FROM AUTHOR]

In recent years, the research on object detection has been intensified. A large number of object detection results are applied to our daily life, which greatly facilitates our work and life. In this paper, we propose a more effective object detection neural network model ENHANCE_YOLOV4. We studied the effects of several attention mechanisms on YOLOV4, and finally concluded that spatial attention mechanism had the best effect on YOLOV4. Therefore, based on previous studies, this paper introduces Dilated Convolution and one-by-one convolution into the spatial attention mechanism to expand the receptive field and combine channel information. Compared with CBAM and BAM, which are composed of spatial attention and channel attention, this improved spatial attention module reduces model parameters and improves detection capabilities. We built a new network model by embedding improved spatial attention module in the appropriate place in YOLOV4. And this paper proves that the detection accuracy of this network structure on the VOC data set is increased by 0.8%, and the detection accuracy on the coco data set is increased by 7%when the calculation performance is increased a little. [ABSTRACT FROM AUTHOR]

In this paper, transfer learning is used for the adaptation of pre-trained deep convolutional neural networks (DCNNs) to find the best appropriate method for the classification of obstructive sleep apnea (OSA) using electrocardiogram (ECG) signals. The Physio Apnea-ECG data set has been used for the evaluation of the proposed method. In deep learning algorithms, especially in image data classification, more data leads to better performance. For this reason, in this paper, we propose a novel technique as follows. First, the ECG signal is divided into 2-s segments and filtered, then the recurrence plots (RP) algorithm is used to convert these segments into two-dimensional images. The RP is an advanced tool that can indicate how resemblances between particular orders vary over time. These plots are generally used for the qualitative evaluation of the time series in dynamic systems. Finally, in the classification stage, 5 pre-trained DCNN models on ImageNet datasets including EfficienNet-B0, EfficienNet-B1, EfficienNet-B2, Inception-v3, and Xception are considered for final decision. By using these methods, the classification accuracies of 88.67%, 90.59%, 90.52%, 93.33%, and 93.19%, were obtained respectively. In this research, by analyzing the performance of the models used, we can see that by increasing the input image size, the number of network parameters, its depth, and the classification performance is improved. Also, the performance of the Inception-v3 model which has the largest input image size and number of parameters with 93.33% accuracy, is better than other models for OSA detection. [ABSTRACT FROM AUTHOR]

Brain abnormality is one of the dominant health issues in recent years, and electroencephalogram (EEG) is a mostly employed method for identifying abnormalities in the brain. Brain abnormalities can be cured only when it is detected early stage. Even though there are various techniques for the detection of brain abnormalities, the health sector concentrates more on the advanced modern method to diagnose the abnormality very precisely. Advanced DL methods have attracted physicians very much due to the fascination with developing detection methods with minimal errors. Hence, this paper deeply concentrated on developing a well-structured technique to predict brain abnormalities using the hybrid coyote predator optimisation-dependent DCNN (hybrid coyote predator-DCNN). The results of this study demonstrate the value of hybrid coyote predator optimisation, which inherits qualities like a long jump and an intelligence ability to adapt to its surroundings and aids in fine-tuning the modal parameters of the DCNN. Additionally, to improve the detection model's precision, the frequency bands from which the features are taken are separated to identify brain changes. Utilising the CAP dataset and the performance metrics, the analysis is carried out. Regarding the TP and k-fold, the hybrid coyote predator-DCNN model attained accuracy of 98.446% and 97.384%, respectively. [ABSTRACT FROM AUTHOR]

Image processing is currently developing as a unique and the inventive field in computer research and applications in the modern area. Most image processing algorithms produce a large quantity of data as an outcome, which is termed as Big-data. These algorithms process and store bulky information either as structured or unstructured data. The use of big data analytics to mine the data produced by image processing technology has huge potential in areas like education, governments, medical establishments, production units, finance and banking, and retail business centers. This paper well defined the innovations made in Big Data analytics and image processing. In this study, a novel data classification approach especially for image analytics is proposed. To improve image quality, pre-processing is applied to huge data that has been gathered. Then, most relevant features like spatial information, texture GLCM, and color and shape features are extracted from the pre-processed image. Since the dimensions of the features are huge in size, an adaptive map-reduce framework with Improved Shannon Entropy has been introduced to lessen the dimensionality of the extracted features. Then, in the big data classification phase, an optimized deep learning classifier deep convolutional neural network (DCNN) is introduced to classify the images accurately. The weight function of the DCNN is fine-tuned using the newly proposed dragonfly updated mothsearch (DAUMS) Algorithm to enhance the classification accuracy and to solve the optimization problems of the research work. The moth search algorithm and dragonfly algorithm are both concepts in this hybrid algorithm DAUMS. [ABSTRACT FROM AUTHOR]

U-Nets occupy the major share in the removal of EMI noises from the CT Images. However, lack of network depth leads to performance degradation in existing variants of U-Nets which remains to be the hardest challenge to remove the EMI noises from the CT scan images. In this paper, a modified U-Nets (URS-Nets++) framework named multi-Headed Attention Based Residual Convolutional Block (MHA-RCB) for the EMI- based Image denoising is proposed. The improvisations in the existing U-Nets are as follows: (1) URS-Nets++ uses the novel convolutional residual networks (CRN) to deepen the network depth, to avoid the network degradation performance. (2) The proposed network introduces the depth-wise attention layers in the convolutional residual network to improve the up-sampling and down-sampling process (3) The proposed URS-Nets++ introduces the modified skip connections (MSP) which are used to fuse the shallow feature information into deeper images details and provides the strong path to obtain the clean images. Extensive experimentation has been conducted for the handcrafted datasets and metrics such as SSIM, PSNR, IoU, DICE are evaluated and analyzed. Ablation study is presented based on original noisy image dataset and standard datasets. Moreover, the performance metrics of the proposed algorithm is compared with the other state-of-the-art deep learning architecture and U-Nets' variants. Results demonstrate that the EMI denoising technique designed by the proposed network yields smoother and sharper images than the other existing methods with better visual quality. [ABSTRACT FROM AUTHOR]

In image classification with Deep Convolutional Neural Networks (DCNNs), the number of parameters in pointwise convolutions rapidly grows due to the multiplication of the number of filters by the number of input channels that come from the previous layer. Existing studies demonstrated that a subnetwork can replace pointwise convolutional layers with significantly fewer parameters and fewer floating-point computations, while maintaining the learning capacity. In this paper, we propose an improved scheme for reducing the complexity of pointwise convolutions in DCNNs for image classification based on interleaved grouped filters without divisibility constraints. The proposed scheme utilizes grouped pointwise convolutions, in which each group processes a fraction of the input channels. It requires a number of channels per group as a hyperparameter C h . The subnetwork of the proposed scheme contains two consecutive convolutional layers K and L, connected by an interleaving layer in the middle, and summed at the end. The number of groups of filters and filters per group for layers K and L is determined by exact divisions of the original number of input channels and filters by C h . If the divisions were not exact, the original layer could not be substituted. In this paper, we refine the previous algorithm so that input channels are replicated and groups can have different numbers of filters to cope with non exact divisibility situations. Thus, the proposed scheme further reduces the number of floating-point computations (11%) and trainable parameters (10%) achieved by the previous method. We tested our optimization on an EfficientNet-B0 as a baseline architecture and made classification tests on the CIFAR-10, Colorectal Cancer Histology, and Malaria datasets. For each dataset, our optimization achieves a saving of 76%, 89%, and 91% of the number of trainable parameters of EfficientNet-B0, while keeping its test classification accuracy. [ABSTRACT FROM AUTHOR]

At present, the image steganography method based on CNN has achieved good results. The trained model and its parameters are of great value. Once leaked, the secret image will be exposed. To protect the security of steganographic network model parameters in the transmission process, an idea based on network model parameter scrambling is proposed in this paper. Firstly, the sender trains the steganography network and extraction network, encrypts the extraction network parameters with the key shared by the sender and the receiver, then sends the extraction network and parameters to the receiver through the public channel, and the receiver recovers them with the key after receiving, to achieve more secure secret communication. In this way, even if the network parameters are intercepted by a third party in the transmission process, the interceptor cannot extract the real secret information. In this paper, the classical Joseph algorithm is used as the scrambling algorithm to scramble the extracted network model parameters of the StegoPNet steganography network. The experimental results show that when the scrambled parameters are used for secret image extraction, a meaningless image independent of the secret image is extracted, it shows that this method can well protect the security of steganography network model. At the same time, this method also has good scalability, and can use a variety of different scrambling algorithms to scramble the parameters. [ABSTRACT FROM AUTHOR]

The remote sensing satellite constellation based on micro-satellites is an important means to construct a global and all-sky earth observation system in the future. Therefore, realizing the recognition of aircraft activities on video micro-satellites is a key technology that needs to be solved urgently. In this paper, an efficient algorithm for aircraft activity recognition that can be deployed on video micro-satellites was proposed. First, aircraft detection was performed on the first incoming remote sensing image using a robust DCNN-based object detection model. Then, a multi-target tracking model incorporating geospatial information was built for aircraft tracking and activity recognition. The algorithm was deployed on an embedded AI computer which was a COTS component. The algorithm was verified using remote sensing videos from commercial micro-satellites. Experimental results show that the algorithm can process aircraft targets of different sizes, and is equally effective even with complex environmental backgrounds, lighting conditions, and various movements of the aircraft, such as turning, entering, and exiting. Based on aircraft tracking results and geospatial information, the motion speed of each aircraft can be obtained, and its activity can be divided into parking, taxiing, or flying. The scheme proposed in this paper has good application prospects in the realization of on-orbit event recognition in micro-satellites with limited computing and memory resources. [ABSTRACT FROM AUTHOR]

The utmost popular dementia cause is because the (AD) Alzheimer's disease. A continuous drop in mental ability is referred to as Dementia. Using the medical images of the brain, the developmental stages of AD symptom of neuropsychiatric functionality are analyzed often. Particularly in the area of classification and detection, cutting edge technologies which comprise computer algorithms, have been used for Alzheimer's disease diagnosis treatments. To improve prediction on new data, ensemble learning applies a group of decision-making systems that uses different strategies to unite classifiers. This paper uses the combination of (DCNN) and deep ensemble learning, i.e., MobileNetV2 and LSTM using magnetic resonance images (MRI). The ADNI dataset is used for the dementia stages classification. Compared to CNN, Deep Ensemble Learning (DEL) performs better. For the evaluation, six metrics were used; accuracy, the area under the curve (AUC), F1- score, precision, recall and computational time. In addition, the calculation for specificity and sensitivity is evaluated for the performance enhancement, which shows the exact affected area. A sensitivity of 94% and a specificity of 95% are obtained in the classification, respectively. [ABSTRACT FROM AUTHOR]

Aquaculture fish diseases pose a serious threat to the security of food supplies. Fish species vary widely, and because they resemble one another so much, it is challenging to distinguish between them based solely on appearance. To stop the spread of disease, it is important to identify sick fish as soon as possible. Due to a lack of necessary infrastructure, it is still difficult to identify infected fish in aquaculture at an early stage. It is essential to promptly identify sick fish to stop the spread of disease. The purpose of this work is to suggest a machine learning technique based on the DCNN method for identifying and categorizing fish diseases. To solve problems involving global optimization, this paper suggests a brand-new hybrid algorithm called the Whale Optimization Algorithm with Genetic Algorithm (WOA-GA) and Ant Colony Optimization. In this work, for classification, the hybrid Random Forest algorithm is used. To facilitate the increased quality, distinctions between both the proposed WOA-GA-based DCNN architecture and the presently used methods for machine learning have been made. The effectiveness of the proposed detection technique is done with MATLAB. Performance metrics like sensitivity, specificity, accuracy, precision, recall, F-measure, NPV, FPR, FNR, and MCC are compared to the performance of the proposed technique. [ABSTRACT FROM AUTHOR]

COVID-19 has engulfed over 200 nations through human-to-human transmission, either directly or indirectly. Reverse Transcription-polymerase Chain Reaction (RT-PCR) has been endorsed as a standard COVID-19 diagnostic procedure but has caveats such as low sensitivity, the need for a skilled workforce, and is time-consuming. Coronaviruses show significant manifestation in Chest X-Ray (CX-Ray) images and, thus, can be a viable option for an alternate COVID-19 diagnostic strategy. An automatic COVID-19 detection system can be developed to detect the disease, thus reducing strain on the healthcare system. This paper discusses a real-time Convolutional Neural Network (CNN) based system for COVID-19 illness prediction from CX-Ray images on the cloud. The implemented CNN model displays exemplary results, with training accuracy being 99.94% and validation accuracy reaching 98.81%. The confusion matrix was utilized to assess the models' outcome and achieved 99% precision, 98% recall, 99% F1 score, 100% training area under the curve (AUC) and 98.3% validation AUC. The same CX-Ray dataset was also employed to predict the COVID-19 disease with deep Convolution Neural Networks (DCNN), such as ResNet50, VGG19, InceptonV3, and Xception. The prediction outcome demonstrated that the present CNN was more capable than the DCNN models. The efficient CNN model was deployed to the Platform as a Service (PaaS) cloud. [ABSTRACT FROM AUTHOR]

As the Internet of Things (IoT) continues to evolve, securing IoT networks and devices remains a continuing challenge. Anomaly detection is a crucial procedure in protecting the IoT. A promising way to perform anomaly detection in the IoT is through the use of machine learning (ML) algorithms. There is a lack of studies in the literature identifying optimal (with regard to both effectiveness and efficiency) anomaly detection models for the IoT. To fill the gap, this work thoroughly investigated the effectiveness and efficiency of IoT anomaly detection enabled by several representative machine learning models, namely Extreme Gradient Boosting (XGBoost), Support Vector Machines (SVMs), and Deep Convolutional Neural Networks (DCNNs). Identifying optimal anomaly detection models for IoT anomaly detection is challenging due to diverse IoT applications and dynamic IoT networking environments. It is of vital importance to evaluate ML-powered anomaly detection models using multiple datasets collected from different environments. We utilized three reputable datasets to benchmark the aforementioned machine learning methods, namely, IoT-23, NSL-KDD, and TON_IoT. Our results show that XGBoost outperformed both the SVM and DCNN, achieving accuracies of up to 99.98%. Moreover, XGBoost proved to be the most computationally efficient method; the model performed 717.75 times faster than the SVM and significantly faster than the DCNN in terms of training times. The research results have been further confirmed by using our real-world IoT data collected from an IoT testbed consisting of physical devices that we recently built. [ABSTRACT FROM AUTHOR]

"MODI lipi" is one of the ancient scripts of Western India. Considerable work has been reported for various other ancient Indian languages except for MODI lipi. Its structural characteristics and non-availability of image database make MODI recognition challenging. The work reported in this paper comprises the creation of an image dataset for MODI handwritten characters and the development of a supervised Transfer Learning (TL)-based classification framework. It makes use of Deep Convolutional Neural Network (DCNN) Alexnet as a pre-trained network to transfer weights to retrain the network. This network is used as a feature extractor to extract features from different layers of the network. A Support Vector Machine (SVM) is trained on activation features to obtain classifier models. These models are investigated further for recognition accuracy and feature analysis. Subjective and objective measures are used to select discriminant deep features. We achieved recognition accuracies of 92.32% and 97.25% for Handwritten MODI character recognition and handwritten Devnagari character recognition, respectively. [ABSTRACT FROM AUTHOR]

The active tracking technology of underwater acoustic targets is an important research direction in the field of underwater acoustic signal processing and sonar, and it has always been issued that draws researchers' attention. The commonly used Kalman filter active tracking (KFAT) method is an effective tracking method, however, it is difficult to detect weak SNR signals, and it is easy to lose the target after the azimuth of different targets overlaps. This paper proposes a KFAT based on deep convolutional neural network (DCNN) method, which can effectively solve the problem of target loss. First, we use Kalman filtering to predict the azimuth and distance of the target, and then use the trained model to identify the azimuth-weighted time-frequency image to obtain the azimuth and label of the target and obtain the target distance by the time the target appears in the time-frequency image. Finally, we associate the data according to the target category, and update the target azimuth and distance information for this cycle. In this paper, two methods, KFAT and DCNN-KFAT, are simulated and tested, and the results are obtained for two cases of tracking weak signal-to-noise signals and tracking different targets with overlapping azimuths. The simulation results show that the DCNN-KFAT method can solve the problem that the KFAT method is difficult to track the target under the weak SNR and the problem that the target is easily lost when two different targets overlap in azimuth. It reduces the deviation range of the active tracking to within 200 m, which is 500~700 m less than the KFAT method. [ABSTRACT FROM AUTHOR]

As an important tool to study the microstructure and properties of materials, High Resolution Transmission Electron Microscope (HRTEM) images can obtain the lattice fringe image (reflecting the crystal plane spacing information), structure image and individual atom image (which reflects the configuration of atoms or atomic groups in crystal structure). Despite the rapid development of HTTEM devices, HRTEM images still have limited achievable resolution for human visual system. With the rapid development of deep learning technology in recent years, researchers are actively exploring the Super-resolution (SR) model based on deep learning, and the model has reached the current best level in various SR benchmarks. Using SR to reconstruct high-resolution HRTEM image is helpful to the material science research. However, there is one core issue that has not been resolved: most of these super-resolution methods require the training data to exist in pairs. In actual scenarios, especially for HRTEM images, there are no corresponding HR images. To reconstruct high quality HRTEM image, a novel Super-Resolution architecture for HRTEM images is proposed in this paper. Borrowing the idea from Dual Regression Networks (DRN), we introduce an additional dual regression structure to ESRGAN, by training the model with unpaired HRTEM images and paired nature images. Results of extensive benchmark experiments demonstrate that the proposed method achieves better performance than the most resent SISR methods with both quantitative and visual results. [ABSTRACT FROM AUTHOR]

The issue of weed classification in paddy fields has received extensive research. There are various methods for resolving the weed classification issue. These methods use color, texture and binary features in classifying the weed image in the paddy field, but it suffers to achieve higher performance. Green–red-texture (GRT) approximation segmentation with the DCNN weed classification model (GRT-DCNN) proposed in this paper to solve this problem. Initially, the method preprocesses the weed image with multiple filters to remove noise from the color image. Further, the process applies K neighbor semicircular color approximation segmentation algorithm toward segmentation. The segmentation scheme groups the weed image features into several groups according to the paddy field nature. With the segmented image, the method applies GRTFA (green–red texture feature approximation) algorithm in selecting the region of interest (ROI). The selected ROI has been used to extract the color features and texture of the selected ROI. The features extracted are used to train the deep convolution neural network. This DCNN model has been designed with four convolution layers to reduce the dimensionality of the features considered. The GR features of the weed image are convolved in two level and texture features are convolved in the next two levels of convolution. The DCNN model was trained with four convolution and pooling layers to perform classification. The neurons at the output layer estimate weed class GR support (WCGRS) toward various classes of weed according to green and color features. Also, the method computes the weed class texture support (WCTS) according to the texture of different weed-maintained classes. The method computes disease class support (DCS) to carry out classification based on the value of WCGRS and WCTS. The decision support system would also estimate the distribution measure according to the growth region to detect the weed growth. The potential yield and the amount of fertilizer needed to support the field worker have been determined. The proposed GRT-DCNN-based weed classification model improves the performance in classification as well as in estimating the yield to support effective regulation of fertilizer. [ABSTRACT FROM AUTHOR]

Protein Remote Homology and fold Recognition (PRHR) is the most crucial task to predict the protein patterns. To achieve this task, Sequence-Order Frequency Matrix-Sampling and Deep learning with Smith-Waterman (SOFM-SDSW) were designed using large-scale Protein Sequences (PSs), which take more time to determine the high-dimensional attributes. Also, it was ineffective since the SW was only applied for local alignment, which cannot find the most matches between the PSs. Hence, in this manuscript, a rapid semi-global alignment algorithm called SOFM-SD-GlobalSW (SOFM-SDGSW) is proposed that facilitates the affine-gap scoring and uses sequence similarity to align the PSs. The major aim of this paper is to enhance the alignment of SW algorithm in both locally and globally for PRHR. In this algorithm, the Maximal Exact Matches (MEMs) are initially obtained by the bit-level parallelism rather than to align the individual characters. After that, a subgroup of MEMs is obtained to determine the global Alignment Score (AS) using the new adaptive programming scheme. Also, the SW local alignment scheme is used to determine the local AS. Then, both local and global ASs are combined to produce a final AS. Further, this resultant AS is considered to train the Support Vector Machine (SVM) classifier to recognize the PRH and folds. Finally, the test results reveal the SOFM-SDGSW algorithm on SCOP 1.53, SCOP 1.67 and Superfamily databases attains an ROC of 0.97, 0.941 and 0.938, respectively, as well as, an ROC50 of 0.819, 0.846 and 0.86, respectively compared to the conventional PRHR algorithms. [ABSTRACT FROM AUTHOR]

Agriculture is a distinct sector of a country's economy. In recent years, new patterns have evolved in the agricultural industry. In conjunction with sensor scaling down and precision agriculture, the field of remote sensor networks, such as the wireless sensor network (WSN), was developed. Its major purpose is to make horticultural operations simpler to identify, assess, and manage. This paper uses the proposed DCNN to predict soil moisture and plan irrigation for precision agriculture farmers to reduce water consumption used for cultivation and increase production yield by comparing water content during various stages of plant growth and integrating IoT applications into agriculture. It also optimizes the water level for future irrigation decisions to maintain crop growth and water stability. The data must be served and stored in the form of a grid view, according to Apriori and GRU (gated recurrent unit). Using numerous sensor and parameter modelling methodologies, this system assists in the prediction of irrigation planning based on irrigation needs. The predicted parameters include soil moisture, temperature, and humidity. This observed experimental data supports smart irrigation in crop production with a high yield and little water use. DCNN has a 98.5% experimental result accuracy rate and the MSE value is predicted in DCNN 99.25% of the time. [ABSTRACT FROM AUTHOR]

In the previous works on heat stress detection, various state-of-the-art machine learning techniques had been used to detect stress patterns from electroencephalographic (EEG) signals. Since the handcrafted feature engineering-based approaches pose certain limitations and are sensitive to transform the nonlinearities of EEG, deep learning techniques have drawn attention in the domain of EEG-based applications. Moreover, existing approaches for stress detection consider the whole frequency band (delta to gamma) which conceals the redundant and lossy information that increased the false detection rate. Also, these approaches were implemented only to identify the stress in binary classes and confined their application to identifying the level of stress. Therefore, a multi-scale channel-wise convolution-based multi-level heat stress assessment has been proposed in this paper. The novel contributions of this work are (1) designing a multi-scale convolutional neural network (MS-CNN) for extracting precise information from individual frequency bands of EEG, (2) use of sparse connectivity to reduce the redundant information and increase the performance with less trainable parameters, and (3) multi-level heat stress assessment for precise identification of the severity of stress. The proposed approaches were evaluated on pre-recorded data of 10 rats in a simulated laboratory environment. The high accuracy of approximately 96–97% and 90–92% has been achieved for binary and multi-level stress detection. There is approximately a 6 to 50% reduction in the trainable parameters with a 2% improvement of accuracy achieved on adopting channel-wise convolution over MS-CNN and deep convolutional neural network (DCNN). This demonstrates the effectiveness of the adopted multiscale feature extraction and sparse connectivity to improve the performance with a less complex model. [ABSTRACT FROM AUTHOR]

The strongly supervised deep convolutional neural network (DCNN) has better performance in assessing breast cancer (BC) because of the more accurate features from the slice-level precise labeling compared with the image-level labeling weakly supervised DCNN. However, manual slice-level precise labeling is time consuming and expensive. In addition, the slice-level diagnosis adopted in the DCNN system is incomplete and defective because of the lack of other slices' information. In this paper, we studied the impact of the region of interest (ROI) and lesion-level multi-slice diagnosis in the DCNN auxiliary diagnosis system. Firstly, we proposed an improved region-growing algorithm to generate slice-level precise ROI. Secondly, we adopted the average weighting method as the lesion-level diagnosis criteria after exploring four different weighting methods. Finally, we proposed our complete system, which combined the densely connected convolutional network (DenseNet) with the slice-level ROI and the average weighting lesion-level diagnosis after evaluating the performance of five DCNNs. The proposed system achieved an AUC of 0.958, an accuracy of 92.5%, a sensitivity of 95.0%, and a specificity of 90.0%. The experimental results showed that our proposed system had a better performance in BC diagnosis because of the more precise ROI and more complete information of multi-slices. [ABSTRACT FROM AUTHOR]

As a potential air control measure, RF-based surveillance is one of the most commonly used unmanned aerial vehicles (UAV) surveillance methods that exploits specific emitter identification (SEI) technology to identify captured RF signal from ground controllers to UAVs. Recently many SEI algorithms based on deep convolution neural network (DCNN) have emerged. However, there is a lack of the implementation of specific hardware. This paper proposes a high-accuracy and power-efficient hardware accelerator using an algorithm–hardware co-design for UAV surveillance. For the algorithm, we propose a scalable SEI neural network with SNR-aware adaptive precision computation. With SNR awareness and precision reconfiguration, it can adaptively switch between DCNN and binary DCNN to cope with low SNR and high SNR tasks, respectively. In addition, a short-time Fourier transform (STFT) reusing DCNN method is proposed to pre-extract feature of UAV signal. For hardware, we designed a SNR sensing engine, denoising engine, and specialized DCNN engine with hybrid-precision convolution and memory access, aiming at SEI acceleration. Finally, we validate the effectiveness of our design on a FPGA, using a public UAV dataset. Compared with a state-of-the-art algorithm, our method can achieve the highest accuracy of 99.3% and an F1 score of 99.3%. Compared with other hardware designs, our accelerator can achieve the highest power efficiency of 40.12 Gops/W and 96.52 Gops/W with INT16 precision and binary precision. [ABSTRACT FROM AUTHOR]

Nowadays, the civil aviation industry has a high precision demand of flight delay prediction. To make full use of the characteristics of flight data and meteorological data, two flight delay prediction models using deep convolution neural network based on fusion of meteorological data are proposed in this paper. One is DCNN (Dual-channel Convolutional Neural Network), which refers to the ResNet network structure. The other is SE-DenseNet (Squeeze and Excitation-Densely Connected Convolutional Network), combining the advantages of DenseNet and SENet. Firstly, flight data and meteorological data are fused in the model. Then, both DCNN and SE-DenseNet models are used to extract feature automatically based on the fused flight data set. Finally, the softmax classifier is adopted to predict the flight delay level. For proposed DCNN model, both straight channel and convolution channel are designed to guarantee the lossless transmission of the feature matrix and enhance the patency of the deep network. For proposed SE-DenseNet model, a SE module is added after the convolution layer of each DenseNet block, which can not only enhance the transmission of deep information but also achieve feature recalibration in the feature extraction process. The research results indicate that after considering characteristics of meteorological information, the accuracy of the model can be improved 1% compared with only considering the flight information. The two deep convolutional neural networks proposed in this paper, DCNN and SE-DenseNet, can both effectively improve the prediction accuracies, reaching to 92.1% and 93.19%, respectively. [ABSTRACT FROM AUTHOR]

In this paper, we investigate the classical detection problem for vehicle networks with multiple antennas, by considering practical communication scenarios, where the interfering signals are correlated over time or frequency. In such cases, the conventional detector requires to estimate the joint distribution of the interfering signals, which imposes a huge computational complexity. To tackle this issue, we propose the joint use of a maximum likelihood detector (MLD) and a deep convolutional neural network (DCNN), where MLD is used to produce an initial detection result and DCNN improves the detection by exploiting the local correlation to suppress the interference. Furthermore, the improved DCNN is enhanced by devising the loss function through the cross-entropy of the detection, which can help to suppress the interfering signals and simultaneously force the residual interference to approach the Gaussian distribution. Simulation results are presented to verify the effectiveness of the proposed detector compared to the conventional one. The trained model and source code for this work are available at https://github.com/skypitcher/project_dcnnmld. [ABSTRACT FROM AUTHOR]

Pedestrian detection using deep convolutional neural networks (DCNNs) has made a breakthrough in the last few years and researchers have proposed different DCNN architectures to detect pedestrians more accurately. Most of these architectures have a backbone based on previous state-of-the-art architectures for classification tasks and just tried to adapt them for their detection task. They are improving their performance with some heuristics, trial and error techniques, and sometimes with grid search on a space of various architectures. However, there is no research in which, firstly, the visual detection system of human has been studied, and then tried to propose a backbone architecture based on that. In this paper, we first review the state-of-the-art methods and then, having a preliminary on visual detection system in the human brain and finally, propose our architecture based on that. The intuition behind our idea can justify the evolutionary course of detection architectures from the first fully convolutional neural networks (FCNNs), like Faster R-CNN, to the modern state-of-the-art methods nowadays and give us a better understanding of why some architectures are superior to the others. The advantage of our idea is that it can be applied to most of the existing architectures with some manipulations, although it is much easier on some methods than others. We have implemented our idea based on an anchor-free method called CSP and could achieve better performance on Caltech-USA and INRIA, which are two of the most popular pedestrian detection datasets. [ABSTRACT FROM AUTHOR]

Diabetic Retinopathy (DR) is a serious consequence of diabetes that seriously impact on the eyes and is a leading cause of blindness. If the lesions in DR arise in the central portion of the fundus, they may result in significant vision loss, which we refer to as Diabetic Macular Edema (DME). Deep learning (DL) techniques are commonly used utilized in ophthalmology for discriminative tasks such as diabetic retinopathy or age-related macular degeneration (AMD) diagnosis. Deep learning techniques typically need huge picture data sets for deep convolutional neural networks (DCNNs) training, it should be graded by human specialists. According to international protocol, it is classified into five severity categories. However, improving a grading model for high generality needs a significant quantity of balanced training data, which is challenging to obtain, especially at high levels of severity. Typical techniques for data augmentation, in many applications of deep learning in the retinal image processing domain, the difficulty of access to huge annotated datasets and legal concerns about patient privacy are limiting issues. As a result, the concept of creating synthetic retinal pictures that are indistinguishable from actual data has garnered more attention. GANs have been certain to be an effective framework for creating synthetic databases of anatomically accurate retinal fundus pictures. GANs, in particular, have garnered increasing attention in ophthalmology. in this article, we present a loss-less generative adversarial network (DR-LL GAN) to generate good resolution fundus pictures that May be adjusted to include random grading and information about the lesion. As a result, large-scale generated data may be used to train a DR grading and lesion segmentation model with more appropriate augmentation. Our model experiments evaluated on IDRID and MESSIDOR datasets, it's obtained a discrimination loss of 0.69374 and a generation loss of 1.10438, as well as a segmentation accuracy of 0.9840 in our tests. This might support in the optimization techniques of the neural network design and in computer-aided screening of medical picture, thus increasing diagnostic reliability for clinical assessment in the future of sophisticated technological healthcare. [ABSTRACT FROM AUTHOR]

Deep convolutional neural networks (DCNNs) have been successfully used in many computer visions task. However, with the increasing complexity of the network and continuous growth of data scale, training a DCNN model suffers from the following three problems: excessive network parameters, insufficient capability of the parameter optimization, and inefficient parallelism. To overcome these obstacles, this paper develops an optimization algorithm for deep convolutional neural network (FP-DCNN) in the MapReduce framework. First, a pruning method based on Taylor's loss (FMPTL) is designed to trim redundant parameters, which not only compresses the structure of DCNN, but also reduces the computational cost of training. Next, a glowworm swarm optimization algorithm based on information sharing strategy (IFAS) is presented, which improves the ability of parameter optimization by adjusting the initialization of weights. Finally, a dynamic load balancing strategy based on parallel computing entropy (DLBPCE) is proposed to achieve an even distribution of data and thus improve the parallel performance of the cluster. Our experiments show that compared with other parallelized algorithms, this algorithm not only reduces the computational cost of network training, but also obtains a higher processing speed. [ABSTRACT FROM AUTHOR]

The Deep Convolutional Neural Network (DCNN) is one of the most powerful and successful deep learning approaches. DCNNs have already provided superior performance in different modalities of medical imaging including breast cancer classification, segmentation, and detection. Breast cancer is one of the most common and dangerous cancers impacting women worldwide. In this paper, we have proposed a method for breast cancer classification with the Inception Recurrent Residual Convolutional Neural Network (IRRCNN) model. The IRRCNN is a powerful DCNN model that combines the strength of the Inception Network (Inception-v4), the Residual Network (ResNet), and the Recurrent Convolutional Neural Network (RCNN). The IRRCNN shows superior performance against equivalent Inception Networks, Residual Networks, and RCNNs for object recognition tasks. In this paper, the IRRCNN approach is applied for breast cancer classification on two publicly available datasets including BreakHis and Breast Cancer (BC) classification challenge 2015. The experimental results are compared against the existing machine learning and deep learning–based approaches with respect to image-based, patch-based, image-level, and patient-level classification. The IRRCNN model provides superior classification performance in terms of sensitivity, area under the curve (AUC), the ROC curve, and global accuracy compared to existing approaches for both datasets. [ABSTRACT FROM AUTHOR]

A key limitation of deep convolutional neural network (DCNN)-based image segmentation methods is the lack of generalizability. Manually traced training images are typically required when segmenting organs in a new imaging modality or from distinct disease cohort. The manual efforts can be alleviated if the manually traced images in one imaging modality (e.g., MRI) are able to train a segmentation network for another imaging modality (e.g., CT). In this paper, we propose an end-to-end synthetic segmentation network (SynSeg-Net) to train a segmentation network for a target imaging modality without having manual labels. SynSeg-Net is trained by using: 1) unpaired intensity images from source and target modalities and 2) manual labels only from source modality. SynSeg-Net is enabled by the recent advances of cycle generative adversarial networks and DCNN. We evaluate the performance of the SynSeg-Net on two experiments: 1) MRI to CT splenomegaly synthetic segmentation for abdominal images and 2) CT to MRI total intracranial volume synthetic segmentation for brain images. The proposed end-to-end approach achieved superior performance to two-stage methods. Moreover, the SynSeg-Net achieved comparable performance to the traditional segmentation network using target modality labels in certain scenarios. The source code of SynSeg-Net is publicly available. 1 https://github.com/MASILab/SynSeg-Net [ABSTRACT FROM AUTHOR]

The study on fault detection and diagnosis (FDD) of chemical processes has always been the top priority of the chemical process safety. In this paper, a fault diagnosis method combining the deep convolutional with the recurrent neural network (DCRNN) is proposed. In this method, the data from chemical processes are input to the deep convolutional neural network (DCNN) to extract features in spatial domains, and then, the features are fused into the bidirectional recurrent neural network (BRNN). Due to the powerful capabilities of DCNN to extract features in spatial domains and the sensitivity to time series of RNN, the combined method can adaptively learn the dynamic information of the raw data in both spatial and temporal domains and has unique advantages in multivariate chemical processes. The application of the DCRNN model in the Tennessee Eastman (TE) process demonstrates the high accuracy of this proposal in identifying abnormal conditions for the chemical process, compared with the traditional fault identification algorithms of deep learning. [ABSTRACT FROM AUTHOR]

In recent years, human activity recognition using radar micro-Doppler signature has been the focus of many researches. Echo signal from body limbs is the origin of motion parameters estimation. To this end, decomposition of this composite signal into components corresponding to each limb is necessary and rather complex. In this paper, a novel method for body micro-Doppler signal decomposition is proposed using a deep convolutional neural network designated DecompNet which uses a context window on the spectral input stream to enhance the decomposition performance. This network is able to decompose 9 components of body limbs. The proposed network, is able to perform signal decomposition and denoising simultaneously, even in low SNR conditions. So, it outperforms the other similar methods both in terms of decomposition ability and robustness to observation noise. Simulation results show that DecompNet can outperform mathematical transform-based methods and also keep working in the SNRs as low as −15dB. [ABSTRACT FROM AUTHOR]

Heart disease is the leading cause of death worldwide. It has an impact on not only the health of patients but also the economies and expenses of the countries. Numerous machine learning and data mining approaches are being developed and explored currently in order to predict various diseases. This paper aims to address the pressing global issue of heart disease by leveraging machine learning and data mining techniques. Specifically, it focuses on utilizing a Fuzzy C means (FCM) approach for attribute segmentation, employing the Whale Optimization Algorithm (WOA) for feature selection, and utilizing Deep Convolutional Neural Networks (DCNNs) for medical diagnosis and early prediction. In this study, the initial stage involves segmenting patient records' attributes using the FCM method. Subsequently, high-ranking features are selected through the WOA algorithm. These segmented features are then input into DCNNs to construct a robust medical diagnosis system and enable early-stage prediction. The DCNNs autonomously extract crucial features without human intervention, enhancing the accuracy of disease prediction. The performance evaluation of the proposed classifier is conducted using the Python platform, with the DCNN achieving an impressive accuracy level of 90.12% during testing. This indicates the DCNN's capability to accurately predict the presence or absence of cardiac disease, showcasing its potential as an effective tool in healthcare. The integration of FCM attribute segmentation, WOA feature selection, and DCNN-based prediction holds significant practical implications. It offers healthcare professionals a valuable tool for diagnosing and predicting heart disease early, potentially saving lives. [ABSTRACT FROM AUTHOR]

Deep convolutional neural networks (DCNNs), with their complex network structure and powerful feature learning and feature expression capabilities, have been remarkable successes in many large-scale recognition tasks. However, with the expectation of memory overhead and response time, along with the increasing scale of data, DCNN faces three non-rival challenges in a big data environment: excessive network parameters, slow convergence, and inefficient parallelism. To tackle these three problems, this paper develops a deep convolutional neural networks optimization algorithm (PDCNNO) in the MapReduce framework. The proposed method first pruned the network to obtain a compressed network in order to effectively reduce redundant parameters. Next, a conjugate gradient method based on modified secant equation (CGMSE) is developed in the Map phase to further accelerate the convergence of the network. Finally, a load balancing strategy based on regulate load rate (LBRLA) is proposed in the Reduce phase to quickly achieve equal grouping of data and thus improving the parallel performance of the system. We compared the PDCNNO algorithm with other algorithms on three datasets, including SVHN, EMNIST Digits, and ISLVRC2012. The experimental results show that our algorithm not only reduces the space and time overhead of network training but also obtains a well-performing speed-up ratio in a big data environment. [ABSTRACT FROM AUTHOR]