Network For Image Recognition Research Articles

A Machine Learning Strategy for Race-Tracking Detection During Manufacturing of Composites by Liquid Moulding

This work presents a supervised machine learning (ML) model to detect race-tracking disturbances during the liquid moulding manufacturing of structural composites. Race-tracking is generated by unexpected resin channels at mould edges that may induce dry spots and porosity formation. The ML model uses the pressure signals recorded by a sensor network as input, providing a classification of the race-tracking event from a set of possible scenarios, and a subsequent variable regression for their position, size and strength. Such a model is based on the residual network (ResNet), a well-known artificial intelligence architecture that makes use of convolutional neural networks for image recognition. Training of the ML classifier and regressors was carried out with the aid of a synthetically generated simulation data set obtained throughout computational fluid dynamics simulations. The time evolution of the pressure sensors was used as grey-level images, or footprints, as inputs to the ResNet ML. The trained model was able to recognise the presence of race-tracking channels from the pressure data yielding good accuracy in terms of label prediction as well as position, size and strength. The model correlation was carried out with a set of injection experiments performed with a constant thickness closed mould containing induced race-tracking channels. The ability of ML models to provide an approximation to the inverse problem, relating the pressure sensor distortions to the cause of such events, is analysed and discussed.

Aircraft Image Recognition Network Based on Hybrid Attention Mechanism.

With the deepening of deep learning research, progress has been made in artificial intelligence. In the process of aircraft classification, the precision rate of aircraft picture recognition based on traditional methods is low due to various types of aircraft, large similarities between different models, and serious texture interference. In this article, the hybrid attention network model (BA-CNN) to implement an aircraft recognition algorithm is proposed to solve the above problems. Using two-channel ResNet-34 as a characteristic extraction function, the depth of network is increased to improve fine-grained characteristic extraction capability without increasing the output characteristic dimension. In the network to introduce a hybrid attention mechanism, respectively, between the residual units of two ResNet-34 channels, channel attention and spatial attention modules are added, more abundant mixed characteristics of attention are obtained, space and characteristics of the local characteristics of the channel response are focused, the characteristics of redundancy are reduced, and the fine-grained characteristics of learning ability are further enhanced. Trained and tested on FGVC-aircraft, a public fine-grained pictures dataset, the recognition precision rate of the BA-CNN networks model reached 89.2%. It can be seen from the experimental results, the recognition precision rate of the original model is improved effectively by using this method, and the recognition precision rate is higher than most of the existing mainstream aircraft recognition ways.

DCCAM-MRNet: Mixed Residual Connection Network with Dilated Convolution and Coordinate Attention Mechanism for Tomato Disease Identification.

Tomato is an important and fragile crop. During the course of its development, it is frequently contaminated with bacteria or viruses. Tomato leaf diseases may be detected quickly and accurately, resulting in increased productivity and quality. Because of the intricate development environment of tomatoes and their inconspicuous disease spot features and small spot area, present machine vision approaches fail to reliably recognize tomato leaves. As a result, this research proposes a novel paradigm for detecting tomato leaf disease. The INLM (integration nonlocal means) filtering algorithm, for example, decreases the interference of surrounding noise on the features. Then, utilizing ResNeXt50 as the backbone, we create DCCAM-MRNet, a novel tomato image recognition network. Dilated Convolution (DC) was employed in STAGE 1 of the DCCAM-MRNet to extend the network's perceptual area and locate the scattered disease spots on tomato leaves. The coordinate attention (CA) mechanism is then introduced to record cross-channel information and direction- and position-sensitive data, allowing the network to more accurately detect localized tomato disease spots. Finally, we offer a mixed residual connection (MRC) technique that combines residual block (RS-Block) and transformed residual block (TR-Block) (TRS-Block). This strategy can increase the network's accuracy while also reducing its size. The DCCAM-classification MRNet's accuracy is 94.3 percent, which is higher than the existing network, and the number of parameters is 0.11 M lesser than the backbone network ResNeXt50, according to the experimental results. As a result, combining INLM and DCCAM-MRNet to identify tomato diseases is a successful strategy.

Batch covariance neural network for image recognition

Recent work has shown that convolutional neural networks (CNN) can achieve state of the art if the datasets are well built. However, the existing convolutional layer is affected by various datasets with the inevitable problems of local abnormal features, i.e., illumination intensity and feature interaction. This paper replaces the convolutional layer with a batch covariance layer (BCL) to locate the category-related region unaffected by the problems. The BCL is regarded as a 3D covariance operation, which calculates the correlation between the kernels and feature maps in kernel size of all channels. Forward propagation, backward propagation, gradient updating, and testing procedure of the BCL are described. The comparison between BCL and convolutional layer shows the ability of BCL to reduce the influence of illumination intensity and feature interaction for discriminating and generating tasks. Complexity analysis shows that BCL can improve the accuracy with a thimbleful time consumption increase. Besides, the batch covariance neural network (BCovNN) is extended from the CNN by replacing the convolutional layer with BCL. Ablation experiment verifies the improvement of BCovNN is provided by BCL separately. BCovNN is evaluated on several popular datasets (i.e., MNIST, STL-10, CIFAR-10, and ImageNet) for image recognition and PASCAL VOC (2007 and 2012) datasets for object localization. Experimental results reveal that BCovNN achieves significant improvements over the corresponding CNN.

Think positive: An interpretable neural network for image recognition

The COVID-19 pandemic is an ongoing pandemic and is placing additional burden on healthcare systems around the world. Timely and effectively detecting the virus can help to reduce the spread of the disease. Although, RT-PCR is still a gold standard for COVID-19 testing, deep learning models to identify the virus from medical images can also be helpful in certain circumstances. In particular, in situations when patients undergo routine X-rays and/or CT-scans tests but within a few days of such tests they develop respiratory complications. Deep learning models can also be used for pre-screening prior to RT-PCR testing. However, the transparency/interpretability of the reasoning process of predictions made by such deep learning models is essential. In this paper, we propose an interpretable deep learning model that uses positive reasoning process to make predictions. We trained and tested our model over the dataset of chest CT-scan images of COVID-19 patients, normal people and pneumonia patients. Our model gives the accuracy, precision, recall and F-score equal to 99.48%, 0.99, 0.99 and 0.99, respectively.

Fault diagnosis of rolling bearing under time-varying speed conditions based on EfficientNetv2

The rolling bearing is a widely-used component in engineering. The fault diagnosis of rolling bearings is key to ensuring the normal operation of equipment. At present, research into the fault diagnosis of rolling bearings mainly focuses on the analysis of vibration data under constant working conditions. However, when dealing with practical engineering problems, equipment frequently operates at variable speed. To analyse the vibration data in the case of frequency conversion and accurately extract the fault characteristic frequency is a challenge, especially when the fault characteristics are weak. In addition, traditional vibration characteristic analysis requires professional technicians to supervise the operation of the equipment, which requires a certain professional ability of the staff. Based on the above two problems, this paper proposes a rolling bearing fault diagnosis model under time-varying speed working conditions, based on the EfficientNetv2 network. This method uses a short-time Fourier transform to convert a one-dimensional vibration signal into a two-dimensional image signal, and uses the advantages of an image recognition network to realize the fault diagnosis under time-varying speed conditions. After training the network, based on transfer learning, the experimental data verify that the accuracy of the results reaches 99.9 ± 0.1%, even in the case of weak fault characteristics, and there is no need for professional technicians to supervise and diagnose once the model is trained, which is conducive to practical application.

Research on Face Intelligent Perception Technology Integrating Deep Learning under Different Illumination Intensities

Aiming at the problem of face recognition under different illumination intensities combined with deep learning algorithm, this research designs a new loss function, i-center loss function and integrates the structure of migration learning algorithm on the basis of LeNets++ deep learning network. The face image data set labled faces in the wild with different illumination intensities and the image data set of supermarket monitoring system are used to train and test the improved LeNets++ deep learning network based on softmax, center and i-center loss function, and a variety of common image recognition networks. The calculation results show that although the amount of data required for the training of LeNets++ deep learning network is much larger than other networks selected in the study, when the loss function is changed to i-center, the accuracy of face image recognition under different light intensities is significantly improved, reaching 99.65%. In the supermarket data set, the maximum face recognition rate of the algorithm using i-center loss function is 99.07%, which is 0.21% and 0.6% higher than that of using center softmax and softmax loss function, respectively. Therefore, experiments show that the improved deep learning neural network based on i-center loss function can improve the effect of face recognition under different illumination intensities.

Federated Resistance Against Adversarial Attacks in Resource-constrained IoT

Federated learning (FL) is a recently evolved distributed learning paradigm that gains increased research attention. To alleviate privacy concerns, FL fundamentally suggests that many entities can cooperatively train the machinedeep learning model by exchanging the learning parameters instead of raw data. Nevertheless, FL still exhibits inherent privacy problems caused by exposing the users’ data based on the training gradients. Besides, the unnoticeable adjustments on inputs done by adversarial attacks pose a critical security threat leading to damaging consequences on FL. To tackle this problem, this study proposes an innovative Federated Deep Resistance (FDR) framework, to provide collaborative resistance against adversarial attacks from various sources in a Fog-assisted IIoT environment. The FDR is designed to enable fog nodes to cooperate to train the FDL model in a way that ensures that contributors have no access to the data of each other, where class probabilities are protected utilizing a private identifier generated for each class. The FDR mainly emphasizes convolutional networks for image recognition from the Food-101 and CIFAR-100 datasets. The empirical results have revealed that FDR outperformed the state-of-the-art adversarial attacks resistance approaches with 5% of accuracy improvements.

Fine-grained Ship Image Recognition Based on BCNN with Inception and燗M-Softmax

The fine-grained ship image recognition task aims to identify various classes of ships. However, small inter-class, large intra-class differences between ships, and lacking of training samples are the reasons that make the task difficult. Therefore, to enhance the accuracy of the fine-grained ship image recognition, we design a fine-grained ship image recognition network based on bilinear convolutional neural network (BCNN) with Inception and additive margin Softmax (AM-Softmax). This network improves the BCNN in two aspects. Firstly, by introducing Inception branches to the BCNN network, it is helpful to enhance the ability of extracting comprehensive features from ships. Secondly, by adding margin values to the decision boundary, the AM-Softmax function can better extend the inter-class differences and reduce the intra-class differences. In addition, as there are few publicly available datasets for fine-grained ship image recognition, we construct a Ship-43 dataset containing 47,300 ship images belonging to 43 categories. Experimental results on the constructed Ship-43 dataset demonstrate that our method can effectively improve the accuracy of ship image recognition, which is 4.08% higher than the BCNN model. Moreover, comparison results on the other three public fine-grained datasets (Cub, Cars, and Aircraft) further validate the effectiveness of the proposed method.

User‐guidedglobal explanations for deep image recognition: A user study

AbstractWe study a user‐guided approach for producing global explanations of deep networks for image recognition. The global explanations are produced with respect to a test data set and give the overall frequency of different “recognition reasons” across the data. Each reason corresponds to a small number of the most significant human‐recognizable visual concepts used by the network. The key challenge is that the visual concepts cannot be predetermined and those concepts will often not correspond to existing vocabulary or have labeled data sets. We address this issue via an interactive‐naming interface, which allows users to freely cluster significant image regions in the data into visually similar concepts. Our main contribution is a user study on two visual recognition tasks. The results show that the participants were able to produce a small number of visual concepts sufficient for explanation and that there was significant agreement among the concepts, and hence global explanations, produced by different participants.

Convolutional Neural Network for Image Recognition

Abstract: Recent developments in the field of machine learning have changed the way it operates for ever, especially with the rise of Artificial Neural Networks (ANN). There is no doubt that these biologically inspired computational models are capable of performing far better than previous forms of artificial intelligence in common machine learning tasks as compared to their previous versions. There are several different forms of artificial neural networks (ANNs), but one of the most impressive is the convolutional neural network (CNN). CNN's have been extensively used for solving difficult pattern recognition tasks using images. With their simple yet precise architecture, they offer a simplified approach to getting started with ANNs. The goal of this paper is to provide a brief introduction to CNN. It discusses the latest papers and newly formed techniques in order to develop these absolutely brilliant models of image recognition. This introduction assumes that you already have a basic understanding of ANNs and machine learning. Keywords: Pattern recognition, artificial neural networks, machine learning, image analysing.

Dialogue Control of Collaborative Robots Based on Artifi cial Neural Networks

Collaborative robotics progress is based on the possibility to apply robots to the wide range activity of peoples. Now the user can control the robot without any special knowledge in robotics and safe. The price of such possibilities is complication of control system of robot which now has to aquire an opportunity of autonomous behavior under human’s control, using the necessary sensors and elements of artificial intelligence. In our research we suppose the collaborative robot as mobile robotic device possible to fulfil some work under the human’s speech demands not only in the same space with the human. We also suppose the necessity of bilateral dialogue human-robot to make it clear the task, the current situation, the state as robot as human. The complex task of control, or may be the collaboration of human with his artificial partner need new means of control, situation recognition, speech dialogue management. As a mean to solve the whole complex of problems we propose the combination of different artificial neural networks. Such as convolution networks for image recognition, deep networks for speech recognition, LSTM networks for autonomous movement of robot control in current situation. Investigations in the field of mobile and manipulation robots including the human-robot control have been proceeded for some years in the department "Robotic systems and mechatronics" BMSTU celebrating now it 70th years Jubilee. The reader may find some of the works in the bibliography. In result of all these investigations we obtain the service robot model which may find a wide application.

Research on Image Recognition System Based on Computer Deep Learning Neural Network

The selection and extraction of image recognition by artificial means needs more complicated work, which is not conducive to the recognition and extraction of important features. Deep learning and neural network represent the iterative expansion of computer intelligent tech, and bring significant results to image recognition. Based on this, this paper first gives the concept and model of neural network, then studies the utilization of deep learning neural network in image recognition, and finally analyses the picture recognition system on account of in-depth learning neural network.

A Modified Gamma Correction Algorithm Applied to the Medium and Bright Light Problems in AI Image Recognition

In this paper, an improved block gamma correction algorithm is proposed for the problem of bright light repair in outdoor scenes in the field of artificial intelligence image recognition, and simulation verification is carried out. Firstly, the strong light area is quickly extracted to meet the high-speed and low-calculation requirements of the artificial intelligence image recognition algorithm. Secondly, the strong light area is divided into sub-blocks, and different gamma values are used according to the standard deviation of each sub-block, which effectively reduces the strong light area. Finally, high image detail information and color confidence are preserved, and at the same time, it meets the requirements of the subsequent image recognition network for the input image.

Prediction of Molecular Properties Using Molecular Topographic Map.

Prediction of molecular properties plays a critical role towards rational drug design. In this study, the Molecular Topographic Map (MTM) is proposed, which is a two-dimensional (2D) map that can be used to represent a molecule. An MTM is generated from the atomic features set of a molecule using generative topographic mapping and is then used as input data for analyzing structure-property/activity relationships. In the visualization and classification of 20 amino acids, differences of the amino acids can be visually confirmed from and revealed by hierarchical clustering with a similarity matrix of their MTMs. The prediction of molecular properties was performed on the basis of convolutional neural networks using MTMs as input data. The performance of the predictive models using MTM was found to be equal to or better than that using Morgan fingerprint or MACCS keys. Furthermore, data augmentation of MTMs using mixup has improved the prediction performance. Since molecules converted to MTMs can be treated like 2D images, they can be easily used with existing neural networks for image recognition and related technologies. MTM can be effectively utilized to predict molecular properties of small molecules to aid drug discovery research.

A Survey of Deep Convolutional Neural Networks Applied for Prediction of Plant Leaf Diseases.

In the modern era, deep learning techniques have emerged as powerful tools in image recognition. Convolutional Neural Networks, one of the deep learning tools, have attained an impressive outcome in this area. Applications such as identifying objects, faces, bones, handwritten digits, and traffic signs signify the importance of Convolutional Neural Networks in the real world. The effectiveness of Convolutional Neural Networks in image recognition motivates the researchers to extend its applications in the field of agriculture for recognition of plant species, yield management, weed detection, soil, and water management, fruit counting, diseases, and pest detection, evaluating the nutrient status of plants, and much more. The availability of voluminous research works in applying deep learning models in agriculture leads to difficulty in selecting a suitable model according to the type of dataset and experimental environment. In this manuscript, the authors present a survey of the existing literature in applying deep Convolutional Neural Networks to predict plant diseases from leaf images. This manuscript presents an exemplary comparison of the pre-processing techniques, Convolutional Neural Network models, frameworks, and optimization techniques applied to detect and classify plant diseases using leaf images as a data set. This manuscript also presents a survey of the datasets and performance metrics used to evaluate the efficacy of models. The manuscript highlights the advantages and disadvantages of different techniques and models proposed in the existing literature. This survey will ease the task of researchers working in the field of applying deep learning techniques for the identification and classification of plant leaf diseases.

Convolutional Neural Network in Image Recognition

Convolutional Neural Network in Image Recognition

Computational Efficiency of a Modular Reservoir Network for Image Recognition.

Liquid state machine (LSM) is a type of recurrent spiking network with a strong relationship to neurophysiology and has achieved great success in time series processing. However, the computational cost of simulations and complex dynamics with time dependency limit the size and functionality of LSMs. This paper presents a large-scale bioinspired LSM with modular topology. We integrate the findings on the visual cortex that specifically designed input synapses can fit the activation of the real cortex and perform the Hough transform, a feature extraction algorithm used in digital image processing, without additional cost. We experimentally verify that such a combination can significantly improve the network functionality. The network performance is evaluated using the MNIST dataset where the image data are encoded into spiking series by Poisson coding. We show that the proposed structure can not only significantly reduce the computational complexity but also achieve higher performance compared to the structure of previous reported networks of a similar size. We also show that the proposed structure has better robustness against system damage than the small-world and random structures. We believe that the proposed computationally efficient method can greatly contribute to future applications of reservoir computing.

Human papilloma virus detection in oropharyngeal carcinomas with in situ hybridisation using hand crafted morphological features and deep central attention residual networks

Human Papilloma Virus (HPV) is a major risk factor for the development of oropharyngeal cancer. Automatic detection of HPV in digitized pathology tissues using in situ hybridisation (ISH) is a difficult task due to the variability and complexity of staining patterns as well as the presence of imaging and staining artefacts. This paper proposes an intelligent image analysis framework to determine HPV status in digitized samples of oropharyngeal cancer tissue micro-arrays (TMA). The proposed pipeline mixes handcrafted feature extraction with a deep learning for epithelial region segmentation as a preliminary step. We apply a deep central attention learning technique to segment epithelial regions and within those assess the presence of regions representing ISH products. We then extract relevant morphological measurements from those regions which are then input into a supervised learning model for the identification of HPV status. The performance of the proposed method has been evaluated on 2009 TMA images of oropharyngeal carcinoma tissues captured with a ×20 objective. The experimental results show that our technique provides around 91% classification accuracy in detecting HPV status when compared with the histopatholgist gold standard. We also tested the performance of end-to-end deep learning classification methods to assess HPV status by learning directly from the original ISH processed images, rather than from the handcrafted features extracted from the segmented images. We examined the performance of sequential convolutional neural networks (CNN) architectures including three popular image recognition networks (VGG-16, ResNet and Inception V3) in their pre-trained and trained from scratch versions, however their highest classification accuracy was inferior (78%) to the hybrid pipeline presented here.

Analysis of the impact of visual attacks on the characteristics of neural networks in image recognition

The article discusses various methods of visual attacks on neural networks such as the introducing of additive white Gaussian noise, pixel maximization and minimization of brightness, maximization and minimization of brightness in the area. For well-known datasets (MNIST, Kaggle dataset called as Dogs_vs_Cats), the dependence of the proportion of correct recognitions on the proportion of distorted images in the test sample is studied. Networks such as VGG-16, Inception_v3, networks without training transfer are analyzed. The characteristics of the accuracy of such networks from the noise level in the image, from the size of the attack areas are obtained. The probabilities of belonging to a given class for various images in various attacks are analyzed. As a method for combating visual attacks, it is proposed to single out a new class during training: "poor image". By selecting valid images by dumping "bad images" by the neural network, accuracy is improved even in the case of a large volume of distorted images in the test sample. The accuracy of such a network is compared for N classes and for (N + 1) classes of images. Ways of further research and ideas of alternative methods of combating visual attacks on neural networks are formulated.