Search

Automatic determination of manufacturing process sequences for the physical production of given part designs is key to facilitate on-demand cyber manufacturing. In this work, we propose an integrated framework that (i) identifies manufacturing features from 3D part designs using a Graph Neural Network (GNN), (ii) identifies the manufacturing processes necessary to produce all features in the part using a Convolutional Neural Network (CNN) that considers shape, material properties, and quality information, and (iii) outputs an ordered manufacturing sequence that can produce the designed part with the help of sequence mining. Using these methods, the knowledge required to enable automated manufacturing process selection is easily scalable and updatable without requiring manual population of ad-hoc or rule-based descriptions. We present exemplar implementations of the proposed framework by suggesting manufacturing sequences for discrete parts with multiple features. The suggested manufacturing sequences demonstrate the potential of the proposed framework for use in future on-demand cyber manufacturing applications. [ABSTRACT FROM AUTHOR]

GoogLeNet is a pre-trained Convolutional Neural Network (CNN) that allows transfer learning and has achieved high recognition rates in image classification tasks. A Recurrence Plot (RP) is an imaging method that depicts the recurrence of the state space system using coloured points and lines in 2D images. This work contributes to facilitating time series feature extraction by proposing a method that applies the GoogLeNet to time series images obtained with RP. The developed method is tested using simulated time series and selected time series from the M3 competition dataset. The results shows that the transfer learning approach allowed the extraction of business time series features by means of a GoogLeNet fine-tuned using 100 simulated time series. The combination of GoogLeNet and RPs outperforms the alternative and easier combination of GoogLeNet and plots of the time series and support the convenience of the RP transformation step. This application of deep learning techniques to business time series imaging offers opportunity for further developments. [ABSTRACT FROM AUTHOR]

Disordered molecular systems, such as amorphous catalysts, organic thin films, electrolyte solutions, and water, are at the cutting edge of computational exploration at present. Traditional simulations of such systems at length scales relevant to experiments in practice require a compromise between model accuracy and quality of sampling. To address this problem, we have developed an approach based on generative machine learning called the Morphological Autoregressive Protocol (MAP), which provides computational access to mesoscale disordered molecular configurations at linear cost at generation for materials in which structural correlations decay sufficiently rapidly. The algorithm is implemented using an augmented PixelCNN deep learning architecture that, as we previously demonstrated, produces excellent results in 2 dimensions (2D) for mono-elemental molecular systems. Here, we extend our implementation to multi-elemental 3D and demonstrate performance using water as our test system in two scenarios: (1) liquid water and (2) samples conditioned on the presence of pre-selected motifs. We trained the model on small-scale samples of liquid water produced using path-integral molecular dynamics simulations, including nuclear quantum effects under ambient conditions. MAP-generated water configurations are shown to accurately reproduce the properties of the training set and to produce stable trajectories when used as initial conditions in quantum dynamics simulations. We expect our approach to perform equally well on other disordered molecular systems in which structural correlations decay sufficiently fast while offering unique advantages in situations when the disorder is quenched rather than equilibrated. [ABSTRACT FROM AUTHOR]

Objective. This study aims to develop and validate an end-to-end software platform, PyHFO, that streamlines the application of deep learning (DL) methodologies in detecting neurophysiological biomarkers for epileptogenic zones from EEG recordings.Approach. We introduced PyHFO, which enables time-efficient high-frequency oscillation (HFO) detection algorithms like short-term energy and Montreal Neurological Institute and Hospital detectors. It incorporates DL models for artifact and HFO with spike classification, designed to operate efficiently on standard computer hardware.Main results. The validation of PyHFO was conducted on three separate datasets: the first comprised solely of grid/strip electrodes, the second a combination of grid/strip and depth electrodes, and the third derived from rodent studies, which sampled the neocortex and hippocampus using depth electrodes. PyHFO demonstrated an ability to handle datasets efficiently, with optimization techniques enabling it to achieve speeds up to 50 times faster than traditional HFO detection applications. Users have the flexibility to employ our pre-trained DL model or use their EEG data for custom model training.Significance. PyHFO successfully bridges the computational challenge faced in applying DL techniques to EEG data analysis in epilepsy studies, presenting a feasible solution for both clinical and research settings. By offering a user-friendly and computationally efficient platform, PyHFO paves the way for broader adoption of advanced EEG data analysis tools in clinical practice and fosters potential for large-scale research collaborations.

DLSIA (Deep Learning for Scientific Image Analysis) is a Python-based machine learning library that empowers scientists and researchers across diverse scientific domains with a range of customizable convolutional neural network (CNN) architectures for a wide variety of tasks in image analysis to be used in downstream data processing. DLSIA features easy-to-use architectures, such as autoencoders, tunable U-Nets and parameter-lean mixed-scale dense networks (MSDNets). Additionally, this article introduces sparse mixed-scale networks (SMSNets), generated using random graphs, sparse connections and dilated convolutions connecting different length scales. For verification, several DLSIA-instantiated networks and training scripts are employed in multiple applications, including inpainting for X-ray scattering data using U-Nets and MSDNets, segmenting 3D fibers in X-ray tomographic reconstructions of concrete using an ensemble of SMSNets, and leveraging autoencoder latent spaces for data compression and clustering. As experimental data continue to grow in scale and complexity, DLSIA provides accessible CNN construction and abstracts CNN complexities, allowing scientists to tailor their machine learning approaches, accelerate discoveries, foster interdisciplinary collaboration and advance research in scientific image analysis.

In recent times, deep learning-based steganalysis classifiers have become popular due to their state-of-the-art performance. Most deep steganalysis classifiers usually extract noise residuals using high-pass filters as preprocessing steps and feed them to their deep model for classification. It is observed that recent steganographic embedding does not always restrict their embedding in the high-frequency zone; instead, they distribute it as per embedding policy. Therefore, besides noise residual, learning the embedding zone is another challenging task. In this work, unlike the conventional approaches, the proposed model first extracts the noise residual using learned denoising kernels to boost the signal-to-noise ratio. After preprocessing, the sparse noise residuals are fed to a novel Multi-Contextual Convolutional Neural Network (M-CNET) that uses heterogeneous context size to learn the sparse and low-amplitude representation of noise residuals. The model performance is further improved by incorporating the Self-Attention module to focus on the areas prone to steganalytic embedding. A set of comprehensive experiments is performed to show the proposed scheme's efficacy over the prior arts. Besides, an ablation study is given to justify the contribution of various modules of the proposed architecture. [ABSTRACT FROM AUTHOR]

Deep learning methods are leveraged in various computer science and artificial intelligence areas, including image classification. Convolutional neural network (CNN) is one of the most widely used deep neural networks for which, several highly effective architectures for image classification have been presented. In this paper, an improved version of the recently introduced CondenseNet is provided as a new network architecture. On the other hand, due to the necessity of reducing the dependence on labeled data in the training process of neural networks, a self-supervised learning method is also proposed for labeling unlabeled images. The results of the experiments show the proper performance of the proposed self-supervised CondenseNet method compared to the basic version of CondenseNet. The experiments are conducted on CIFAR_10 and CIFAR-100 datasets and show better accuracy of the proposed method. [ABSTRACT FROM AUTHOR]

High-resolution remote sensing images (RSI) refer to images captured from a distance, usually from an aircraft or satellite that provide details about the Earth's surface. It can be used in several application areas like environmental monitoring, urban planning, agriculture, and disaster response. In urban planning, high-resolution imagery is used to monitor the growth of urban areas or to recognize the area that requires infrastructure improvement. Vehicle detection is a significant way of understanding high-resolution RSIs. Vehicle detection and classification on high-resolution RSI is a difficult task that needs a group of computer vision (CV), image processing, and machine learning (ML) algorithms. Deep convolutional neural network (DCNN) based techniques have attained recent outcomes in many object detection datasets and have enriched several CV tasks. This article designed and developed a sparrow search optimization algorithm with deep learning for vehicle type detection and classification (SSOADL-VTDC) technique on high-resolution remote sensing images. The presented SSOADL-VTDC technique examines the high-quality RSIs for the accurate detection and classification of vehicles. To accomplish this, the SSOADL-VTDC technique employs a YOLOv5 object detector with a Residual Network as a backbone approach. In addition, the SSOADL-VTDC technique uses SSOA based hyperparameter optimizer designed for the parameter tuning of the YOLOv5 model. For the vehicle classification process, the SSOADL-VTDC technique exploits the softmax classifier. The simulation validation of the SSOADL-VTDC approach was validated on a high-resolution RSI dataset and the outcomes demonstrated the greater of the SSOADL-VTDC methodology in terms of different measures. [ABSTRACT FROM AUTHOR]

Human Activity Recognition (HAR) deals with the automatic recognition of physical activities and plays a crucial role in healthcare and sports where wearable sensors and intelligent computational techniques are used. We propose a HAR algorithm that uses the smartphones accelerometer data for human activity recognition. In particular, we present a recurrent convolutional neural network-based HAR algorithm that combines a Convolutional Neural Network (CNN) to extract temporal features from the sensor data, a Fuzzy C-Means (FCM) clustering algorithm to cluster the features extracted by the CNN, and a Long Short-Term Memory (LSTM) network to learn the temporal dependencies between the features. We evaluate the proposed methodology on two distinct datasets: the MotionSense dataset and the WISDM dataset. We evaluate the proposed CNN-FCM-LSTM model on the publicly available MotionSense dataset to classify ten activity types: 1) walking upstairs, 2) walking downstairs, 3) jogging, 4) sitting, 5) standing, 6) level ground walking, 7) jumping jacks, 8) brushing teeth, 9) writing, and 10) eating. Next, we evaluate the model's performance on the WISDM dataset to assess its ability to generalize to unseen data. On the MotionSense test dataset, CNN-FCM-LSTM achieves a classification accuracy of 99.69%, a sensitivity of 99.62%, a specificity of 99.63%, and a false positive rate per hour (FPR/h) of 0.37%. Meanwhile, it achieves a classification accuracy of 97.27% on the WISDM dataset. The CNN-FCM-LSTM model's capability to classify a diverse range of activities within a single architecture is noteworthy. The results suggest that the proposed CNN-FCM-LSTM model using smartphone inputs is more accurate, reliable, and robust in detecting and classifying activities than the state-of-the-art models. It should be noted that activity recognition technology has the potential to aid in studying the underpinnings of physical activity, designing more effective training regimens, and simulating the rigors of competition in sports. [ABSTRACT FROM AUTHOR]

Sign language serves as an alternative communication mode for those with limitations in hearing and speaking abilities. In India, an estimated population of around 2.7 million people have hearing or speech impairments. Among this population, a significant majority of 98% use sign language as their primary mode of communication. Unfortunately, the limited availability of human interpreters is a considerable obstacle in recognizing and identifying diverse sign languages. To tackle this issue, the present study aims to undertake a thorough examination of various computational methodologies used in various geographical areas for the purpose of categorizing and identifying discrete sign languages. Among the pool of 587 research papers deemed appropriate for qualitative assessment, a total of 95 studies were specifically focused on the categorization and identification of sign language using artificial intelligence-based techniques. The study examines prior research involving deep learning and machine learning methods for a systematic literature review of Sign Language Recognition (SLR). The categorization is made based on language, and the study investigates several facets, including sign type, signing modes, processing techniques, classification methodologies, and evaluation measures. The study reveals that extensive studies were carried out in Chinese, Arabic, and American sign languages. The findings of this review show that among the machine learning techniques, Support Vector Machine (SVM) exhibited higher performance measures, while Convolutional Neural Networks (CNN) exhibited higher performance among the deep learning techniques. [ABSTRACT FROM AUTHOR]

For the safety and reliability of the system, Remaining Useful Life (RUL) prediction is considered in many industries. The traditional machine learning techniques must provide more feature representation and adaptive feature extraction. Deep learning techniques like Long Short-Term Memory (LSTM) achieved an excellent performance for RUL prediction. However, the LSTM network mainly relies on the past few data, which may only capture some contextual information. This paper proposes a hybrid combination of Convolution Neural Network (CNN) and LSTM (CNN+LSTM) to solve this problem. The proposed hybrid model predicts how long a machine can operate without breaking down. In the proposed work, 1D horizontal and vertical signals of the mechanical bearing are first converted to 2D images using Continuous Wavelet Transform (CWT). These 2D images are applied to CNN for key feature extraction. Ultimately, these key features are applied to the LSTM deep neural network for predicting the RUL of a mechanical bearing. A PRONOSTIA data is utilized to demonstrate the performance of the proposed model and compare the proposed model with other state-of-the-art methods. Experimental results show that our proposed CNN+LSTM-based hybrid model achieved higher accuracy (98%) with better robustness than existing methods. [ABSTRACT FROM AUTHOR]

Convolutional neural networks (CNNs) have demonstrated promising accuracy in segmenting concrete cracks under controlled conditions. However, these existing methods usually are challenging in addressing multi-scale patterns or complex scenarios. To address this pervasive issue, a modified U-Net network was established in the current study. The proposed model, namely U-Net-ASPP-CBAM, integrates the convolutional block attention module after a convolution operator to selectively focus on the crack information while disregarding irrelevant background details during feature extraction. Furthermore, U-Net-ASPP-CBAM network replaces a pooling layer using the atrous spatial pyramid pooling module to explore and fuse features across multiple scales, enhancing information capture for the segmentation of small objects. The performance of the proposed model has been validated by a self-built dataset comprising crack images with diverse complex backgrounds. And the segmentation effectiveness is assessed through evaluation indices, including precision, recall, F1 score, pixel accuracy, and mean intersection over union. The results show the proposed U-Net-ASPP-CBAM model outperforms other segmentation models. [ABSTRACT FROM AUTHOR]

Camouflaged object detection (COD) intends to find concealed objects hidden in the surroundings. COD is challenging for it has to discriminate the minor difference between foreground and background. In most existing methods, convolutional neural network (CNN)-based approaches are proposed to overcome this challenge. However, they have limitations in extracting semantic features of input images and learning global contexts. This study presents a novel Transformer and CNN-based Depth Context Aggregation network (call DCA-Net) for concealed object detection and segmentation. This network uses Swin Transformer as backbone to extract globalized semantic features. Dilated Reception (DR) module is designed to connect the encoder and decoder. A hybrid loss function is used for optimizing the model. In particular, to supplement the intersection over union (IOU) loss, Complementary (C) loss is introduced. By analyzing different metrics, comprehensive comparisons with previous methods are conducted on four public datasets, such as CAMO, CHAMELEON, COD10K, and NC4K. Experimental analysis proves that the proposed DCA-Net achieves state-of-the-art performance. [ABSTRACT FROM AUTHOR]

Many people are affected by lung cancer and it kills a lot of people worldwide. Often times, lung cancer does not have symptoms in the early stage. Abnormal growths called pulmonary nodules may be found in the lungs. Some of the pulmonary nodules may be cancerous. So, identifying pulmonary nodules will help detect lung cancer at an early stage. V-3DResNets based on variants of residual network is proposed to improve the generalization capability of the pulmonary nodule detection system. A U-shaped encoder-decoder is used to effectively extract the features. Residual networks help to reduce the vanishing gradient problem in deep neural networks. A Slice Grouped Non Local (SGNL) module studies the relation among entities across Computed Tomography (CT) slices. The content-aware mechanism of the squeeze and excitation block assigns different weights to the channels and improves the performance of the nodule detection system. The proposed work is analyzed using the PN9 dataset that includes 8,798 CT Scans with nine types of pulmonary nodules which attained a Competition Performance metric (CPM) of 65.51%. [ABSTRACT FROM AUTHOR]

For Deepfake detection, many existing works use the cross-entropy loss to enforce the classifier network to learn the mapping relationship from the RGB domain to the class domain, lacking an explicit constraint to guide the feature extraction network to learn discriminative features from an input image. This constrains the feature representation capability to expose deepfake. In this work, we analyze the feature extraction network in terms of both difference and similarity capabilities and propose a new constraint called similarity loss (SL) to improve the detection performance of the convolutional neural network (CNN) based detector. Moreover, according to the experimental results of the SL on data augmentation effectiveness, we propose a simple yet efficient framework, which is called as efficient similarity representation learning (ESRL), for deepfake detection. Extensive experiments on three public datasets (namely FF++, DFDC, and Celeb-DF) show that the feature extraction network trained with the help of SL can map forged faces and real faces to different feature embedding and map the same type of forged faces to similar feature embedding. [ABSTRACT FROM AUTHOR]

Airway diseases cause significant challenges while diagnosing it accurately as well as timely. In current years, the combination of deep learning techniques with Explainable Artificial Intelligence (XAI), has emerged as a promising possibility to enhance the understanding as well as interpretability of complex medical data which are associated with these conditions. The study explores the usage of advanced convolutional neural networks (CNNs) on large-scale of Chest X-ray or CT scan image datasets and their application in the domain of medical imaging to detect as well as classify airway diseases such as pneumoconiosis, pulmonary embolism, Covid-19, tuberculosis, and Lung cancer. The visualization of the images have been enhanced by applying sharpening techniques and are later subsequently used to train various deep learning models which include InceptionResNetV2, MobileNet, Xception, ResNet152, EfficientNetV2M, and DenseNet169. Further, various parameters have been used for the evaluation of the performance of these models in determining the most optimal choice. During training the model, it has been found that the highest accuracy, precision, recall, and F1 score has been computed by InceptionResNetV2 with 99.15%, 98% and 99% (both) respectively on a loss of 0.0275. Besides this, a graphical user interface (GUI) has been also developed to improvise the interaction between the AI based model and end users. The GUI is based on the theory of XAI which after predicting the diseases provides in detail explanation of treatment as well as preventive measures of the disease. The motto of the research is to escalate the knowledge of deep learning techniques for various lung based issues as well as to enlighten with the significance of XAI in promoting interpretability as well as transparency. The findings of the research implies towards the advancement of decision support systems at various medical sectors to promote effective as well as personalized interventions in the realm of pulmonology. [ABSTRACT FROM AUTHOR]

Speech emotion recognition holds significant importance as it enables machines to understand and respond to human emotions, enhancing human-computer interaction and personalized experiences. Accurate identification and interpretation of emotional states from speech signals enable various benefits, including enhanced personalized experiences, effective monitoring of mental health, and improved human-computer interfaces. However, recognizing emotions from speech is a difficult task primarily because there exists a significant disparity between acoustic features and human emotions. Both vocal cues and spoken words play significant roles in determining a person's emotional state. Therefore, in order to accurately identify human emotions from speech, it is essential to extract distinct and meaningful acoustic features. In this paper, we propose a novel approach to infer human emotional states. Human emotional state recognition has a wide range of applications ranging from customer service to mental health. Our proposed approach extracts a set of features from the speech signals, and employs a framework known as deep stride convolutional neural network using bi-directional LSTM. Our proposed model achieved a high accuracy of 95% which is almost 20% higher when compared to the state of the art on the RAVDESS dataset, while also minimizing the loss. [ABSTRACT FROM AUTHOR]

Brain–computer interfacing (BCI) research, fueled by deep learning, integrates insights from diverse domains. A notable focus is on steady-state visual evoked potential (SSVEP) in BCI applications, requiring in-depth assessment through deep learning. EEG research frequently employs SSVEPs, which are regarded as normal brain responses to visual stimuli, particularly in investigations of visual perception and attention. This paper tries to give an in-depth analysis of the implications of deep learning for SSVEP-adapted BCI. A systematic search across four stable databases (Web of Science, PubMed, ScienceDirect, and IEEE) was developed to assemble a vast reservoir of relevant theoretical and scientific knowledge. A comprehensive search yielded 177 papers that appeared between 2010 and 2023. Thence a strict screening method from predetermined inclusion criteria finally generated 39 records. These selected works were the basis of the study, presenting alternate views, obstacles, limitations and interesting ideas. By providing a systematic presentation of the material, it has made a key scholarly contribution. It focuses on the technical aspects of SSVEP-based BCI, EEG technologies and complex applications of deep learning technology in these areas. The study delivers more penetrating reporting on the latest deep learning pattern recognition techniques than its predecessors, together with progress in data acquisition and recording means suitable for SSVEP-based BCI devices. Especially in the realms of deep learning technology orchestration, pattern recognition techniques, and EEG data collection, it has effectively closed four important research gaps. To increase the accessibility of this critical material, the results of the study take the form of easy-to-read tables just generated. Applying deep learning techniques in SSVEP-based BCI applications, as the research shows, also has its downsides. The study concludes that a radical framework will be presented which, includes intelligent decision-making tools for evaluation and benchmarking. Rather than just finding a comparable or similar analogy, this framework is intended to help guide future research and pragmatic applications, and to determine which SSVEP-based BCI applications have succeeded at responsibility for what they set out with. [ABSTRACT FROM AUTHOR]

Polyp segmentation in colorectal images is the most effective and necessary tool for the early detection of colorectal cancer, and deep learning has become popular for efficiently segmenting polyps. The complex morphological characteristics of polyps, such as the unclear boundary between polyps and mucosa, and the lack of training data could cause great difficulties in network fitting. Transformer-based semantic segmentation networks have achieved more promising performance than traditional convolutional neural networks. However, the dispersion of self-attention and the less accurate local feature recognition limit the further development and applications of Transformer-based networks. This paper proposes a novel Stepwise Attention Emphasis module to refocus self-attention for Transformer-based polyp segmentation in colorectal images, where a reverse fuse module is used to better fuse different levels of features. Furthermore, a new decoder network, called the densely smooth fusion decoder, is also proposed to enhance local details and provide more useful information from deep features to shallow features. Experimental comparisons are conducted, and result analysis shows that the proposed network achieves promising performance in both learning and generalization ability on public datasets. [ABSTRACT FROM AUTHOR]

Several types of cancer can develop in different organs at the same time, and lung and colon cancer can affect the human body adversely in some cases. To diagnose cancer cases, experts must go through a lengthy and complicated process of analyzing histopathological images. Still, the process can be quickly achieved with the technological possibilities currently available. However, lung and colon cancer (LCC) classification requires improvement. For this reason, we developed a method for classifying LCC using a deep learning technique. In the proposed methodology, there are three phases, namely pre-processing, segmentation, and classification. As a first step, histopathological images are collected and preprocessed. It is necessary to use a median filter for pre-processing. The segmentation process is then conducted on the pre-processed images. The swin transformer is used to segment the data. The lung nodules and colon are segmented as a result of the segmentation process. To classify the image as normal or cancerous, these segmented parts are fed into the enhanced cascade convolution neural network classifier (EC2N2). To improve the efficiency of the cascade classifier, the hyper-parameters are optimally selected using the Adaptive Tasmanian Devil Optimization (ATDO) algorithm. The efficiency of the presentedtechnique is analysed based on various metrics and research work compared with different state-of-art-works. [ABSTRACT FROM AUTHOR]

Skin cancer is a disorder that is becoming more prevalent around the world and is responsible for numerous mortality. Skin cancer starts in one organ and slowly moves to other parts of the body before killing the patient. Early detection of skin cancer is important for reducing the number of deaths all over the world. Because it takes time and money to manually diagnose skin cancer, it is critical to create automated diagnostic techniques to categorize skin lesions more accurately. Medical image enhancement and deep learning-based segmentation techniques are developed in this proposed work. Cancer affected and non-affected skin images are given as input for the proposed method. Data collection consists of raw data that cannot produce high accuracy. So, a certain pre-processing technique is used in the proposed method to achieve high accuracy. Dingo Optimized Texture based Histogram Equalization (DOTHE) strategy is utilized to improve the skin image. Then the pre-processed image is partitioned into different parts or regions according to the features and properties of the pixels in the image. U-Net network architecture is used in the proposed method to segment the enhanced image. The performances of the proposed model are analyzed using the Convolutional Neural Network (CNN) model. This proposed model is tested with several metrics which attain better performances like 97% accuracy, 96% sensitivity, 95% specificity, 94% precision, and 3% error. Thus the designed model enhances and segments the image effectively, and it is useful for effective skin cancer prediction. [ABSTRACT FROM AUTHOR]

The sixth generation (6G) networks demand high security, low latency, and highly dependable standards and capacity. One of the essential components of 6G networks is flexible wireless network slicing. In this paper, we propose a hybrid model that combines a convolutional neural network (CNN) and a bidirectional long short-term memory (BiLSTM). The hybrid model is applied to the Unicauca IP Flow Version2 dataset. The CNN handles the automated feature section, while the BiLSTM is utilized for categorizing the suitable network slices. This hybrid model is capable of offering a reliable and effective network slice to the end user. The proposed hybrid model has an overall recognition rate of 97.21%, which reflects the applicability of the proposed approach. A stratified 10-fold cross-validation is used to assess the applicability of the proposed model. The main challenge for network service providers is to assign slices correctly. A clever method is needed to make a standard for accurately assigning network slices to an unidentified device when it asks for them. For each incoming request for new traffic, the proposed model forecasts the suitable network slice [ABSTRACT FROM AUTHOR]

Blockchain technology has unlocked the doors of building decentralized applications, where security plays a vital role. Any transaction ever created in a blockchain is recorded permanently. With the increasing usage of social media modalities today, many non-reputable sources create and publish fake and luring news. Due to the easy availability of the internet and excessive use of social media, fake news can spread like a flash. It causes the need to eliminate fake news posted on social media platforms. The main motive of the paper is to develop an ensemble model that assists in the automated identification of fake news. This paper addresses the identified gaps in detecting multi-modal fake news. We propose Fake News Network (FNNet) as a novel secure blockchain-based deep learning model to detect fake news. It uses blockchain and deep learning to assure data integrity and learn data representations, respectively. The proposed model consists of four layers: the node layer, the deep learning layer, the blockchain layer, and the network layer. This is an ensemble deep-learning model that leverages Bi-LSTM and CNN models, where the Bi-LSTM model captures the sequential data in both directions, and CNN captures the hidden features. We use Pheme, CrisisLexT6, and ISOT datasets to train the model. FNNet achieves an average accuracy of 86.93%, 92.22%, and 98.53% on Pheme, CrisisLexT9, and ISOT datasets, respectively. Our results show that the proposed model is robust and applicable to real-time datasets and social media networks to detect fake news effectively. [ABSTRACT FROM AUTHOR]

The accurate health status evaluation of lithium-ion batteries is crucial for preemptive identification of potential battery failures and averting hazardous incidents, given its essential role in indicating the extent of battery degradation. The challenge in determining the State of Health (SOH) arises from the absence of a precise and standardized definition, as well as the difficulty in measuring essential input variables. Therefore, this paper utilizes current and voltage data during the charge and discharge process as direct inputs for SOH estimation and proposes a deep learning-based lithium-ion battery SOH estimation approach. Specifically, it leverages Bayesian optimized Convolutional Neural Network (CNN) within a data-driven framework. Experimental results demonstrate that the proposed deep learning method achieves a Mean Absolute Error (MAE) of 1% and a Maximum Error (MAX) below 4% in estimation accuracy, highlighting its enhanced precision and robustness. [ABSTRACT FROM AUTHOR]

In recent years, video surveillance has become an integral part of computer vision research, addressing a variety of challenges in security, memory management and content extraction from video sequences. This paper introduces the Robust Object Detection using Fire Hawks Optimizer with Deep Learning (ROD-FHODL) technique, a novel approach designed specifically for video surveillance applications. Combining object detection and classification the proposed technique employs a two-step procedure. Utilizing the power of the Mask Region-based Convolutional Neural Network (Mask-RCNN) for object detection, we optimize its hyperparameters using the Fire Hawks Optimizer (FHO) algorithm to improve its efficacy. Our experimental results on the UCSD dataset demonstrate the significant impact of the proposed work. It achieves an extraordinary RUNNT of 1.34 s on the pedestrian-1 dataset, significantly outperforming existing models. In addition, the proposed system surpasses in accuracy, with a pedestrian-1 accuracy rate of 97.45% and Area Under the Curve (AUC) values of 98.92%. Comparative analysis demonstrates the superiority of the proposed system in True Positive Rate (TPR) versus False Positive Rate (FPR) across thresholds. In conclusion, the proposed system represents a significant advancement in video surveillance, offering advances in speed, precision and robustness that hold promise for enhancing security, traffic management and public space monitoring in smart city infrastructure and other applications. [ABSTRACT FROM AUTHOR]

Switchgear is a vital component of modern power systems, responsible for regulating the flow of electrical power. Contact issues, irregular loads, and other similar problems can cause switchgear to overheat, leading to unexpected disturbances and potential damage to the power equipment. Thermal imaging shows significant potential and is increasingly employed to detect faults in power equipment. However, the unique characteristics of thermal images often pose challenges to accurate fault detection. This research aims to study the effectiveness of transfer learning architectures for switchgear fault detection. This paper applies eleven transfer learning architectures, including SqueezeNet, GoogLeNet, InceptionV3, DenseNet201, ResNet50, Xception, InceptionResNetV2, ShuffleNet, EfficientNetB0, AlexNet, and VGG19. The results of the testing phase demonstrated that the application of transfer learning by fine-tuning pre-trained convolutional neural network architectures was highly effective in the classification of thermal images captured from switchgear units. The models achieved accuracy rates between 83.87% and 98.38%, and values of F 1 -Scores between 83.11% and 98.34% in the pre-trained architectures. [Display omitted] • Transfer learning is applied to address the unique characteristics of thermal images for switchgear fault detection. • A thermal image dataset is setup and augmented to improve the classification performance. • Establish standard comparisons among eleven transfer learning architectures. • Achieve rapid and precise classification with the proposed transfer learning architectures. [ABSTRACT FROM AUTHOR]

The flood-prone city of Zaio in Morocco necessitates a precise, fast, real-time flood depth mapping model due to its recurrent flood history. Whether it's the exclusive prediction of one flood category, relying on hard-to-measure inputs like flood hydrographs, or employing less accurate training methods such as cellular automata models, the existing Convolutional Neural Network (CNN) models face limitations in predicting flood depth in a city whose flood dynamics are influenced by outer watersheds such as Zaio. This study addresses these issues by introducing a CNN tailored for real-time pluvial and fluvial flood depth mapping in Zaio at fine resolution (2 m). Training involved eight rainfall hyetographs, with four used for testing. The model's validation included three "unseen" rainfall distribution patterns. The proposed CNN seamlessly connects rainfall-runoff modeling and hydrodynamic 2D flood depth simulation, with a novelty of predicting both pluvial and fluvial flood depth, and reducing computational time by a significant 99.17%. [ABSTRACT FROM AUTHOR]

To maintain the security of transformer differential protection, it is essential to restrain its response to oversetting differential current caused by the inrush current or other switching conditions. This paper presents a new proposed method to discriminate the transformer's internal fault from the inrush current; the discrimination process is based on a convolutional neural network (CNN) with a combination of the higher order spectral estimations that perform a deep learning classification with high accuracy. This research succeeded in proposing two robust and efficient CNN models; the first one is the 1D CNN, which takes the sole signal without any transformation, while the second model is the 2D CNN, which takes the short‐time Fourier transform of the signal. Both developed models are light and have the minimum number of layers that can achieve a very high performance. The performance of the proposed models is tested using a dataset prepared according to laboratory‐measured transformer parameters. The results are compared with other well‐known methods, and the achievement of high numerical performance evaluation validates the consistency of the proposed methods. [ABSTRACT FROM AUTHOR]

ABSTRACT Introduction Methods Results Conclusion VEXAS is a syndrome described in 2020, caused by mutations of the UBA1 gene, and displaying a large pleomorphic array of clinical and hematological features. Nevertheless, these criteria lack significance to discriminate VEXAS from other inflammatory conditions at the screening step. This work hence first focused on singling out dysplastic features indicative of the syndrome among peripheral blood (PB) polymorphonuclears (PMN). A deep learning algorithm is then proposed for automatic detection of these features.A multicentric dataset, comprising 9514 annotated PMN images was gathered, including UBA1 mutated VEXAS (n = 25), UBA1 wildtype myelodysplastic (n = 14), and UBA1 wildtype cytopenic patients (n = 25). Statistical analysis on a subset of patients was performed to screen for significant abnormalities. Detection of these features on PB was then automated with a convolutional neural network (CNN) for multilabel classification.Significant differences were observed in the proportions of PMNs with pseudo‐Pelger, nuclear spikes, vacuoles, and hypogranularity between patients with VEXAS and both cytopenic and myelodysplastic controls.Automatic detection of these abnormalities yielded AUCs in the range [0.85–0.97] and a F1‐score of 0.70 on the test set. A VEXAS screening score was proposed, leveraging the model outputs and predicting the UBA1 mutational status with 0.82 sensitivity and 0.71 specificity on the test patients.This study suggests that computer‐assisted analysis of PB smears, focusing on suspected VEXAS cases, can provide valuable insights for determining which patients should undergo molecular testing. The presented deep learning approach can help hematologists direct their suspicions before initiating further analyses. [ABSTRACT FROM AUTHOR]

Background Methods Results Discussion Deep learning methods are revolutionizing natural science. In this study, we aim to apply such techniques to develop blood type prediction models based on cheap to analyze and easily scalable screening array genotyping platforms.Combining existing blood types from blood banks and imputed screening array genotypes for ~111,000 Danish and 1168 Finnish blood donors, we used deep learning techniques to train and validate blood type prediction models for 36 antigens in 15 blood group systems. To account for missing genotypes a denoising autoencoder initial step was utilized, followed by a convolutional neural network blood type classifier.Two thirds of the trained blood type prediction models demonstrated an F1‐accuracy above 99%. Models for antigens with low or high frequencies like, for example, Cw, low training cohorts like, for example, Cob, or very complicated genetic underpinning like, for example, RhD, proved to be more challenging for high accuracy (>99%) DL modeling. However, in the Danish cohort only 4 out of 36 models (Cob, Cw, D‐weak, Kpa) failed to achieve a prediction F1‐accuracy above 97%. This high predictive performance was replicated in the Finnish cohort.High accuracy in a variety of blood groups proves viability of deep learning‐based blood type prediction using array chip genotypes, even in blood groups with nontrivial genetic underpinnings. These techniques are suitable for aiding in identifying blood donors with rare blood types by greatly narrowing down the potential pool of candidate donors before clinical grade confirmation. [ABSTRACT FROM AUTHOR]

Background: Regional Wall Motion Abnormality (RWMA) serves as an early indicator of myocardial infarction (MI), the global leader in mortality. Accurate and early detection of RWMA is vital for the successful treatment of MI. Current automated echocardiography analyses typically concentrate on peak values from left ventricular (LV) displacement curves, based on LV contour annotations or key frames during the heart's systolic or diastolic phases within a single echocardiographic cycle. This approach may overlook the rich motion field features available in multi-cycle cardiac data, which could enhance RWMA detection. Methods: In this research, we put forward an innovative approach to detect RWMA by harnessing motion information across multiple echocardiographic cycles and multi-views. Our methodology synergizes U-Net-based segmentation with optical flow algorithms for detailed cardiac structure delineation, and Temporal Convolutional Networks (ConvNet) to extract nuanced motion features. We utilize a variety of machine learning and deep learning classifiers on both A2C and A4C views echocardiograms to enhance detection accuracy. A three-phase algorithm—originating from the HMC-QU dataset—incorporates U-Net for segmentation, followed by optical flow for cardiac wall motion field features. Temporal ConvNet, inspired by the Temporal Segment Network (TSN), is then applied to interpret these motion field features, independent of traditional cardiac parameter curves or specific key phase frame inputs. Results: Employing five-fold cross-validation, our SVM classifier demonstrated high performance, with a sensitivity of 93.13%, specificity of 83.61%, precision of 88.52%, and an F1 score of 90.39%. When compared with other studies using the HMC-QU datasets, these Fig s stand out, underlining our method's effectiveness. The classifier also attained an overall accuracy of 89.25% and Area Under the Curve (AUC) of 95%, reinforcing its potential for reliable RWMA detection in echocardiographic analysis. Conclusions: This research not only demonstrates a novel technique but also contributes a more comprehensive and precise tool for early myocardial infarction diagnosis. [ABSTRACT FROM AUTHOR]

Background: Mild cognitive impairment (MCI) is the transition stage between the cognitive decline expected in normal aging and more severe cognitive decline such as dementia. The early diagnosis of MCI plays an important role in human healthcare. Current methods of MCI detection include cognitive tests to screen for executive function impairments, possibly followed by neuroimaging tests. However, these methods are expensive and time-consuming. Several studies have demonstrated that MCI and dementia can be detected by machine learning technologies from different modality data. This study proposes a multi-stream convolutional neural network (MCNN) model to predict MCI from face videos. Results: The total effective data are 48 facial videos from 45 participants, including 35 videos from normal cognitive participants and 13 videos from MCI participants. The videos are divided into several segments. Then, the MCNN captures the latent facial spatial features and facial dynamic features of each segment and classifies the segment as MCI or normal. Finally, the aggregation stage produces the final detection results of the input video. We evaluate 27 MCNN model combinations including three ResNet architectures, three optimizers, and three activation functions. The experimental results showed that the ResNet-50 backbone with Swish activation function and Ranger optimizer produces the best results with an F1-score of 89% at the segment level. However, the ResNet-18 backbone with Swish and Ranger achieves the F1-score of 100% at the participant level. Conclusions: This study presents an efficient new method for predicting MCI from facial videos. Studies have shown that MCI can be detected from facial videos, and facial data can be used as a biomarker for MCI. This approach is very promising for developing accurate models for screening MCI through facial data. It demonstrates that automated, non-invasive, and inexpensive MCI screening methods are feasible and do not require highly subjective paper-and-pencil questionnaires. Evaluation of 27 model combinations also found that ResNet-50 with Swish is more stable for different optimizers. Such results provide directions for hyperparameter tuning to further improve MCI predictions. [ABSTRACT FROM AUTHOR]

Cracks exist in the majority of components of ancient Chinese timber structures and have led to serious mechanical property degradation problems, threatening the safety of the whole structures and making the cracks' detection for wooden components a necessity. With the rapid development of intelligent protection technology of cultural buildings, it is important to establish a scientific identification and quantification method for cracks in wooden components to replace traditional manual detection techniques. Deep learning is precisely such an advanced technology. In this study, images of cracked wooden components were first collected from the Yingxian wooden pagoda and the crack characteristics were analyzed. A dataset for crack segmentation was established using a total of 501 images of cracked wooden components, including a training dataset of 450 images and a validation dataset of 51 images. Based on the mathematical principles of deep learning and the fully convolutional neural networks (FCNN), a deep fully convolutional neural network (d‐FCNN) model was constructed based on encoding and decoding methodology. Four model indicators, pixel accuracy (PA), average pixel accuracy (mPA), mean intersection over union (mIoU), and F1‐score were analyzed to train the model and determine the optimal model parameters, including learning rate, batch size, and epoch. It concluded that the optimal initial learning rate takes the value of 10−4, batch size of 6, and epoch of 100, achieving the average accuracy of 78.8%. Further, based on the pixel's accumulation principle, a quantitative calculation method for crack length and maximum width was proposed. Two cracked wooden columns were prepared, and crack image identification and quantification experiments were conducted to verify the correctness of the constructed d‐FCNN model and the proposed crack quantification method. The results show that the model is suitable for crack intelligence detection, identification, and quantification of cracked wooden components. [ABSTRACT FROM AUTHOR]

Nowadays, transfer learning has shown promising results in many applications. However, most deep transfer learning methods such as parameter sharing and fine‐tuning are still suffering from the lack of parameters transmission strategy. In this paper, we propose a new optimization model for parameter‐based transfer learning in convolutional neural networks named STP‐CNN. Indeed, we propose a Lasso transfer model supported by a regularization term that controls transferability. Moreover, we opt for the proximal gradient descent method to solve the proposed model. The suggested technique allows, under certain conditions, to control exactly which parameters, in each convolutional layer of the source network, which will be used directly or adjusted in the target network. Several experiments prove the performance of our model in locating the transferable parameters as well as improving the data classification. [ABSTRACT FROM AUTHOR]

This work aims to explore the application of an improved convolutional neural network (CNN) combined with Internet of Things (IoT) technology in art design education and teaching. The development of IoT technology has created new opportunities for art design education, while deep learning and improved CNN models can provide more accurate and effective tools for image processing and analysis. In order to enhance the effectiveness of art design teaching and students' creative expression, this work proposes an improved CNN model. In model construction, it increases the number of convolutional layers and neurons, and incorporates the batch normalization layer and dropout layer to enhance feature extraction capabilities and reduce overfitting. Besides, this work creates an experimental environment using IoT technology, capturing art image samples and environmental data using cameras, sensors, and other devices. In the model application phase, image samples undergo preprocessing and are input into the CNN for feature extraction. Sensor data are concatenated with image feature vectors and input into the fully connected layers to comprehensively understand the artwork. Finally, this work trains the model using techniques such as cross-entropy loss functions and L2 regularization and adjusts hyperparameters to optimize model performance. The results indicate that the improved CNN model can effectively acquire art sample data and student creative expression data, providing accurate and timely feedback and guidance for art design education and teaching, with promising applications. This work offers new insights and methods for the development of art design education. [ABSTRACT FROM AUTHOR]

Alzheimer's disease (AD), the most prevalent form of dementia, requires early prediction for timely intervention. Current deep learning approaches, particularly those using traditional neural networks, face challenges such as handling high-dimensional data, interpreting complex relationships, and managing data bias. To address these limitations, we propose a framework utilizing graph neural networks (GNNs), which excel in modeling relationships within graph-structured data. Our study employs GNNs on data from the Alzheimer's Disease Neuroimaging Initiative for binary and multi-class classification across the three stages of AD: cognitively normal (CN), mild cognitive impairment (MCI), and Alzheimer's disease (AD). By incorporating comorbidity data derived from electronic health records, we achieved the most effective multi-classification results. Notably, the GNN model (Chebyshev Convolutional Neural Networks) demonstrated superior performance with a 0.98 accuracy in multi-class classification and 0.99, 0.93, and 0.94 in the AD/CN, AD/MCI, and CN/MCI binary tasks, respectively. The model's robustness was further validated using the Australian Imaging, Biomarker & Lifestyle dataset as an external validation set. This work contributes to the field by offering a robust, accurate, and cost-effective method for early AD prediction (CN vs. MCI), addressing key challenges in existing deep learning approaches. [ABSTRACT FROM AUTHOR]

This study constructed deep learning models using plain skull radiograph images to predict the accurate postnatal age of infants under 12 months. Utilizing the results of the trained deep learning models, it aimed to evaluate the feasibility of employing major changes visible in skull X-ray images for assessing postnatal cranial development through gradient-weighted class activation mapping. We developed DenseNet-121 and EfficientNet-v2-M convolutional neural network models to analyze 4933 skull X-ray images collected from 1343 infants. Notably, allowing for a ± 1 month error margin, DenseNet-121 reached a maximum corrected accuracy of 79.4% for anteroposterior (AP) views (average: 78.0 ± 1.5%) and 84.2% for lateral views (average: 81.1 ± 2.9%). EfficientNet-v2-M reached a maximum corrected accuracy 79.1% for AP views (average: 77.0 ± 2.3%) and 87.3% for lateral views (average: 85.1 ± 2.5%). Saliency maps identified critical discriminative areas in skull radiographs, including the coronal, sagittal, and metopic sutures in AP skull X-ray images, and the lambdoid suture and cortical bone density in lateral images, marking them as indicators for evaluating cranial development. These findings highlight the precision of deep learning in estimating infant age through non-invasive methods, offering the progress for clinical diagnostics and developmental assessment tools. [ABSTRACT FROM AUTHOR]

Molecular subtypes of colorectal cancer (CRC) significantly influence treatment decisions. While convolutional neural networks (CNNs) have recently been introduced for automated CRC subtype identification using H&E stained histopathological images, the correlation between CRC subtype genomic variants and their corresponding cellular morphology expressed by their imaging phenotypes is yet to be fully explored. The goal of this study was to determine such correlations by incorporating genomic variants in CNN models for CRC subtype classification from H&E images. We utilized the publicly available TCGA-CRC-DX dataset, which comprises whole slide images from 360 CRC-diagnosed patients (260 for training and 100 for testing). This dataset also provides information on CRC subtype classifications and genomic variations. We trained CNN models for CRC subtype classification that account for potential correlation between genomic variations within CRC subtypes and their corresponding cellular morphology patterns. We assessed the interplay between CRC subtypes' genomic variations and cellular morphology patterns by evaluating the CRC subtype classification accuracy of the different models in a stratified 5-fold cross-validation experimental setup using the area under the ROC curve (AUROC) and average precision (AP) as the performance metrics. The CNN models that account for potential correlation between genomic variations within CRC subtypes and their cellular morphology pattern achieved superior accuracy compared to the baseline CNN classification model that does not account for genomic variations when using either single-nucleotide-polymorphism (SNP) molecular features (AUROC: 0.824±0.02 vs. 0.761±0.04, p<0.05, AP: 0.652±0.06 vs. 0.58±0.08) or CpG-Island methylation phenotype (CIMP) molecular features (AUROC: 0.834±0.01 vs. 0.787±0.03, p<0.05, AP: 0.687±0.02 vs. 0.64±0.05). Combining the CNN models account for variations in CIMP and SNP further improved classification accuracy (AUROC: 0.847±0.01 vs. 0.787±0.03, p = 0.01, AP: 0.68±0.02 vs. 0.64±0.05). The improved accuracy of CNN models for CRC subtype classification that account for potential correlation between genomic variations within CRC subtypes and their corresponding cellular morphology as expressed by H&E imaging phenotypes may elucidate the biological cues impacting cancer histopathological imaging phenotypes. Moreover, considering CRC subtypes genomic variations has the potential to improve the accuracy of deep-learning models in discerning cancer subtype from histopathological imaging data. [ABSTRACT FROM AUTHOR]

Summary Due to the rise in information rate prerequisite and the heterogeneity level, the modification in network traffic in the upcoming wireless communication (WC) encompasses innovative challenges in the case of energy efficiency (EE) and spectrum management. To tackle this issue, several existing techniques have been imposed but none of the frameworks provided effective solutions to compatible with recent WC applications. This framework introduces an innovative deep learning (DL)–based distributed cognitive radio network (DCRN). The proposed scheme emphasizes single base station (BS) management, where resource effectiveness is obtained by solving active resource allocation (RA) problems using a bipartite matching (BM) technique. A DL scheme is emphasized to predict the traffic load (TL) for effective EE using a residual inception‐enriched recurrent convolutional neural network (R‐InceptionRCNN). The proposed method is implemented in Python, and the performance metrics including uplink (UL) achievable capacity per secondary user (SU), UL achievable capacity per SU, cost of energy consumption, EE, and mean energy saving (MES) are scrutinized and compared with conventional techniques. The proposed scheme achieved the overall costs, EE, MES, and UL capacity of 14.33 C/J, 149.99 J/MB, 13.49%, and 22.33 Mbps, respectively, on performing RA and TL prediction in the CRN platform. [ABSTRACT FROM AUTHOR]

ABSTRACT Tree‐ring data are pivotal for decoding the age and growth patterns of trees, reflecting the impact of environmental factors over time. Addressing the significant shortcomings of traditional, labour‐intensive and resource‐demanding methods, we propose an innovative automated technique that utilizes panchromatic images and deep learning for measuring tree rings. The method utilizes convolutional neural networks to enhance image quality, precisely delineate tree rings through segmentation and perform ring counting and width calculation in the post‐processing stage. We compiled an extensive data set from diverse sources, including Beijing Forestry University and the Summer Palace, to train our algorithm. The performance of our method was validated empirically, demonstrating its potential to transform tree‐ring analysis and provide deeper insights into ecological and climatological research. [ABSTRACT FROM AUTHOR]

Background Purpose Methods Results Conclusions Endoscopic instrument segmentation is essential for ensuring the safety of robotic‐assisted spinal endoscopic surgeries. However, due to the narrow operative region, intricate surrounding tissues, and limited visibility, achieving instrument segmentation within the endoscopic view remains challenging.This work aims to devise a method to segment surgical instruments in endoscopic video. By designing an endoscopic image classification model, features of frames before and after the video are extracted to achieve continuous and precise segmentation of instruments in endoscopic videos.Deep learning techniques serve as the algorithmic core for constructing the convolutional neural network proposed in this study. The method comprises dual stages: image classification and instrument segmentation. MobileViT is employed for image classification, enabling the extraction of key features of different instruments and generating classification results. DeepLabv3+ is utilized for instrument segmentation. By training on distinct instruments separately, corresponding model parameters are obtained. Lastly, a flag caching mechanism along with a blur detection module is designed to effectively utilize the image features in consecutive frames. By incorporating specific parameters into the segmentation model, better segmentation of surgical instruments can be achieved in endoscopic videos.The classification and segmentation models are evaluated on an endoscopic image dataset. In the dataset used for instrument segmentation, the training set consists of 7456 images, the validation set consists of 829 images, and the test set consists of 921 images. In the dataset used for image classification, the training set consists of 2400 images and the validation set consists of 600 images. The image classification model achieves an accuracy of 70% on the validation set. For the segmentation model, experiments are conducted on two common surgical instruments, and the mean Intersection over Union (mIoU) exceeds 98%. Furthermore, the proposed video segmentation method is tested using videos collected during surgeries, validating the effectiveness of the flag caching mechanism and blur detection module.Experimental results on the dataset demonstrate that the dual‐stage video processing method excels in performing instrument segmentation tasks under endoscopic conditions. This advancement is significant for enhancing the intelligence level of robotic‐assisted spinal endoscopic surgeries. [ABSTRACT FROM AUTHOR]

A novel image‐driven deep learning approach embedded with model‐data‐knowledge information can directly predict the full field stress distribution and mechanical properties of composites solely based on initial scanning geometry images. The fiber spatial distribution and morphological features are identified by Scanning Electron Microscopy (SEM) initially. An effective random fiber generation algorithm is further utilized to generate images database comprising equivalent geometric models of various microstructures. Subsequently, the geometry model images database is analyzed by fast Fourier transform (FFT) to produce the database including the elastic modulus parameters and full field stress distributions of composites with distinct microstructures. Finally, an innovative image‐learning comprehensive paradigm uniting convolutional neural networks and convolutional autoencoders is elaborated systematically to learn the inherent laws of geometry images and field images. The results show that the proposed method have capacity to effectively and accurately predict the elastic modulus and full field stress distributions combined with error analysis even for stochastic boundary composites. The proposed methodology is a symbiosis of cutting‐edge image learning and non‐destructive testing techniques for composites, which can directly use the local non‐destructive or scanning images of composite structural components to quickly obtain field information and further predict damage evolution online. Highlights: Proposing a novel deep learning approach combined with FFT directly to predict stress distribution of stochastic boundary composites solely based on micro‐images.Adopting equivalent geometric models dispersed by higher pixels mesh density considering extreme thin interface.Incorporate adaptive algorithms for streamlined training optimization.An innovative integration of deep learning and non‐destructive testing technology for the stress distribution and properties prediction online. [ABSTRACT FROM AUTHOR]