Image Recognition Problems Research Articles

Research on low illumination coal gangue image enhancement based on improved Retinex algorithm

ABSTRACT Aiming at the problems of coal and gangue image recognition in low illumination or high dust concentration environment, an adaptive enhancement method based on improved Retinex algorithm, is proposed: H-GF-MSR algorithm. The algorithm is processed in HSV color space, uses guided filter for multi-scale Retinex algorithm in brightness component, and carries out adaptive saturation stretching in saturation component. Then the images are processed by histogram equalization, and the two algorithms are fused to improve the accuracy. Compared with SSR algorithm, MSRCR algorithm and other classical algorithms, this algorithm is improved in suppressing the generation of halo and image edge enhancement. The experimental results show that in terms of image evaluation indexes such as information entropy, image mean, image average gradient and peak signal-to-noise ratio, compared with SSR algorithm, the algorithm improves by 5.4%, 41.1%, 33.2% and 0.4% respectively; compared with MSRCR algorithm, the average improvement is 2.9%, 32.4%, 31.7% and 0.7% respectively. Through the enhancement experiments of coal gangue image in the illumination range of 10 lux ~ 200 lux, the algorithm has better enhancement effect. It can effectively suppress the bright light and restore the original details of the object.

Image Classification and Recognition Based on Deep Learning and Random Forest Algorithm

Convolutional neural network (CNN) is an important way to solve the problems of image classification and recognition. It can realize effective feature representation and make continuous breakthroughs in the field of image recognition, but it needs a lot of time in the training process. At the same time, random forest (RF) has the advantages of fast training speed and high classification accuracy. Aiming at the problem of image classification and recognition, this paper proposes a hybrid model based on CNN, which inputs the features extracted by CNN into RF for classification. Since the random weight network can also obtain valid results, the gradient algorithm is not used to adjust the network parameters to avoid consuming a lot of time. Finally, experiments are conducted on MNIST dataset and rotated MNIST dataset, and the results show that the classification accuracy of the hybrid model has improved more than RF, and also, the generalization ability has been improved.

Classification of Underwater Fish Images and Videos via Very Small Convolutional Neural Networks

The automatic classification of fish species appearing in images and videos from underwater cameras is a challenging task, albeit one with a large potential impact in environment conservation, marine fauna health assessment, and fishing policy. Deep neural network models, such as convolutional neural networks, are a popular solution to image recognition problems. However, such models typically require very large datasets to train millions of model parameters. Because underwater fish image and video datasets are scarce, non-uniform, and often extremely unbalanced, deep neural networks may be inadequately trained, and undergo a much larger risk of overfitting. In this paper, we propose small convolutional neural networks as a practical engineering solution that helps tackle fish image classification. The concept of “small” refers to the number of parameters of the resulting models: smaller models are lighter to run on low-power devices, and drain fewer resources per execution. This is especially relevant for fish recognition systems that run unattended on offshore platforms, often on embedded hardware. Here, established deep neural network models would require too many computational resources. We show that even networks with little more than 12,000 parameters provide an acceptable working degree of accuracy in the classification task (almost 42% for six fish species), even when trained on small and unbalanced datasets. If the fish images come from videos, we augment the data via a low-complexity object tracking algorithm, increasing the accuracy to almost 49% for six fish species. We tested the networks with images obtained from the deployments of an experimental system in the Mediterranean sea, showing a good level of accuracy given the low quality of the dataset.

SELECTING BIVARIATE COPULA MODELS USING IMAGE RECOGNITION

AbstractThe choice of a copula model from limited data is a hard but important task. Motivated by the visual patterns that different copula models produce in smoothed density heatmaps, we consider copula model selection as an image recognition problem. We extract image features from heatmaps using the pre-trained AlexNet and present workflows for model selection that combine image features with statistical information. We employ dimension reduction via Principal Component and Linear Discriminant Analyses and use a Support Vector Machine classifier. Simulation studies show that the use of image data improves the accuracy of the copula model selection task, particularly in scenarios where sample sizes and correlations are low. This finding indicates that transfer learning can support statistical procedures of model selection. We demonstrate application of the proposed approach to the joint modelling of weekly returns of the MSCI and RISX indices.

Detecting autism from picture book narratives using deep neural utterance embeddings.

Deficits in the ability to use language in social contexts, including storytelling skills, are observed across the autism spectrum. Development in machine-learning approaches may contribute to clinical psychology and psychiatry, given its potential to support decisions concerning the diagnosis and treatment of psychiatric conditions and disorders. To evaluate the usefulness of deep neural networks for detecting autism spectrum disorder (ASD) from textual utterances, specifically from narrations produced by individuals with ASD. We examined two text encoders: Embeddings from Language Models (ELMo) and Universal Sentence Encoder (USE), and three classification algorithms: XGBoost, support vector machines, and dense neural network layer. We aimed to classify 25 participants with ASD and 25 participants with typical development (TD) based on their narrations produced during the Autism Diagnostic Observation Schedule, Second Edition (ADOS-2) picture book task. The results of computational approaches were compared with the results of standardized testing and classifications made by two psychiatrists (raters). The raters were asked to read utterances produced by a participant (without an examiner's statements and additional information) and assign a participant to one of the two groups: ASD or with typical development (TD). The computer-based models had higher sensitivity, specificity, positive predictive values and negative predictive values than the raters, and lower than the two standardized instruments: ADOS-2 and Social Communication Questionnaire (SCQ). Our findings lay the groundwork for future studies involving deep neural network-based text representation models as tools for augmenting the ASD diagnosis or screening. Both ELMo and USE text encoders provided promising specificities, sensitivities, positive predictive values and negative predictive values. Our results indicate the usefulness of page-level embeddings for utterance representation in ADOS-2 picture book task. What is already known on this subject Deficits in the use of language in social contexts, and narrative ability in particular, are observed across the autism spectrum. Most research on narrative skills has applied hand-coding methods. Hitherto, machine-learning methods were used mostly for image recognition problems and data from screening questionnaires for ASD classification. Detection of mental and developmental disorders from textual input is an emerging field for machine and deep-learning methods. What this paper adds to existing knowledge This study explored the ability of several types of deep neural network-based text representation models to detect ASD. Both ELMo and USE provided the most promising values of specificity, sensitivity, positive predictive values and negative predictive values. What are the potential or actual clinical implications of this work? Competitive accuracy, repeatability, speed and ease of operation are all advantages of computerized methods. They allow for objective and quantitative assessment of narrative ability and complex language skills. Deep neural network-based text representation models could in the future support clinicians and augment the decision-making process related to ASD diagnosis, screening and intervention planning.

Symmetric Binary Tree Based Co-occurrence Texture Pattern Mining for Fine-grained Plant Leaf Image Retrieval

Leaf image patterns have been actively researched for plant species recognition. However, as a very challenging fine-grained pattern identification issue, cultivar recognition in which the leaf image patterns usually have very subtle difference among cultivars has not yet received considerable attention in computer vision and pattern recognition community. In this paper, a novel symmetric geometric configuration, named Symmetric Binary Tree (SBT) which has multiple symmetric branch pairs and can change in size, is designed to mine the multiple scale co-occurrence texture patterns. The resulting SBT descriptors encode both shape and texture features which make them more informative than the existing individual descriptors and co-occurrence features. A novel feature fusion scheme, named K-NN Based Handcrafted and Deep Features Fusion (KNN-HDFF) that encodes the neighbouring information of distance measure, is proposed for further boosting the retrieval performance. Extensive experiments conducted on the challenging soybean cultivar leaf image dataset and peanut cultivar leaf image dataset consistently indicate the superiority of the proposed method over the state-of-the-art methods on fine-grained leaf image retrieval. We also conduct extensive experiments of feature fusions using the proposed KNN-HDFF on the benchmark datasets and the experimental results prove its potential for improving the performance of cultivar identification which also indicates that fusing handcrafted and deep features may be the direction to address the challenging fine-grained image recognition problem.

DeepRepeat: direct quantification of short tandem repeats on signal data from nanopore sequencing

Despite recent improvements in basecalling accuracy, nanopore sequencing still has higher error rates on short-tandem repeats (STRs). Instead of using basecalled reads, we developed DeepRepeat which converts ionic current signals into red-green-blue channels, thus transforming the repeat detection problem into an image recognition problem. DeepRepeat identifies and accurately quantifies telomeric repeats in the CHM13 cell line and achieves higher accuracy in quantifying repeats in long STRs than competing methods. We also evaluate DeepRepeat on genome-wide or candidate region datasets from seven different sources. In summary, DeepRepeat enables accurate quantification of long STRs and complements existing methods relying on basecalled reads.

High Similarity Image Recognition and Classification Algorithm Based on Convolutional Neural Network.

Nowadays, the information processing capabilities and resource storage capabilities of computers have been greatly improved, which also provides support for the neural network technology. Convolutional neural networks have good characterization capabilities in computer vision tasks, such as image recognition technology. Aiming at the problem of high similarity image recognition and classification in a specific field, this paper proposes a high similarity image recognition and classification algorithm fused with convolutional neural networks. First, we extract the image texture features, train different types, and different resolution image sets and determine the optimal texture different parameter values. Second, we decompose the image into subimages according to the texture difference, extract the energy features of each subimage, and perform classification. Then, the input image feature vector is converted into a one-dimensional vector through the alternating 5-layer convolution and 3-layer pooling of convolutional neural networks. On this basis, different sizes of convolution kernels are used to extract different convolutions of the image features, and then use convolution to achieve the feature fusion of different dimensional convolutions. Finally, through the increase in the number of training and the increase in the amount of data, the network parameters are continuously optimized to improve the classification accuracy in the training set and in the test set. The actual accuracy of the weights is verified, and the convolutional neural network model with the highest classification accuracy is obtained. In the experiment, two image data sets of gems and apples are selected as the experimental data to classify and identify gems and determine the origin of apples. The experimental results show that the average identification accuracy of the algorithm is more than 90%.

Explaining classifiers by constructing familiar concepts

Interpreting a large number of neurons in deep learning is difficult. Our proposed ‘CLAssifier-DECoder’ architecture (ClaDec) facilitates the understanding of the output of an arbitrary layer of neurons or subsets thereof. It uses a decoder that transforms the incomprehensible representation of the given neurons to a representation that is more similar to the domain a human is familiar with. In an image recognition problem, one can recognize what information (or concepts) a layer maintains by contrasting reconstructed images of ClaDec with those of a conventional auto-encoder(AE) serving as reference. An extension of ClaDec allows trading comprehensibility and fidelity. We evaluate our approach for image classification using convolutional neural networks. We show that reconstructed visualizations using encodings from a classifier capture more relevant classification information than conventional AEs. This holds although AEs contain more information on the original input. Our user study highlights that even non-experts can identify a diverse set of concepts contained in images that are relevant (or irrelevant) for the classifier. We also compare against saliency based methods that focus on pixel relevance rather than concepts. We show that ClaDec tends to highlight more relevant input areas to classification though outcomes depend on classifier architecture. Code is at https://github.com/JohnTailor/ClaDec

A Convolutional Neural Network-Based Recognition Method of Gear Performance Degradation Mode

Abstract In an increasingly intelligent modern society, whether in industrial production activities or daily life, mechanical transmission equipment is more and more widely used. Once a failure occurs, it will not only cause the stagnation of industrial production, bring huge economic losses and environmental pollution, but may also cause casualties. Therefore, it is particularly important to identify and monitor the performance degradation of mechanical equipment. Based on the convolutional neural network (CNN), a stacking incremental deformable residual block network recognition model is proposed. This method converts the one-dimensional signal recognition problem into an image recognition problem. The average pooling layer replaces the fully connected layer, and the large-size convolution kernel is replaced with a small-size convolution kernel. With the recognition of the gear performance degradation modes, the experiment proves that the multi-channel recognition model has a better recognition effect.

Accuracy and Resiliency of Analog Compute-in-Memory Inference Engines

Recently, analog compute-in-memory (CIM) architectures based on emerging analog non-volatile memory (NVM) technologies have been explored for deep neural networks (DNNs) to improve scalability, speed, and energy efficiency. Such architectures, however, leverage charge conservation, an operation with infinite resolution, and thus are susceptible to errors. Thus, the inherent stochasticity in any analog NVM used to execute DNNs, will compromise performance. Several reports have demonstrated the use of analog NVM for CIM in a limited scale. It is unclear whether the uncertainties in computations will prohibit large-scale DNNs. To explore this critical issue of scalability, this article first presents a simulation framework to evaluate the feasibility of large-scale DNNs based on CIM architecture and analog NVM. Simulation results show that DNNs trained for high-precision digital computing engines are not resilient against the uncertainty of the analog NVM devices. To avoid such catastrophic failures, this article introduces the analog bi-scale representation for the DNN, and the Hessian-aware Stochastic Gradient Descent training algorithm to enhance the inference accuracy of trained DNNs. As a result of such enhancements, DNNs such as Wide ResNets for CIFAR-100 image recognition problem are demonstrated to have significant performance improvements in accuracy without adding cost to the inference hardware .

Discriminative information restoration and extraction for weakly supervised low-resolution fine-grained image recognition

The existing methods of fine-grained image recognition mainly devote to learning subtle yet discriminative features from the high-resolution input. However, their performance deteriorates significantly when they are used for low quality images because a lot of discriminative details of images are missing. We propose a discriminative information restoration and extraction network, termed as DRE-Net, to address the problem of low-resolution fine-grained image recognition, which has widespread application potential, such as shelf auditing and surveillance scenarios. DRE-Net is the first framework for weakly supervised low-resolution fine-grained image recognition and consists of two sub-networks: (1) fine-grained discriminative information restoration sub-network (FDR) and (2) recognition sub-network with the semantic relation distillation loss (SRD-loss). The first module utilizes the structural characteristic of minimum spanning tree (MST) to establish context information for each pixel by employing the spatial structures between each pixel and other pixels, which can help FDR focus on and restore the critical texture details. The second module employs the SRD-loss to calibrate recognition sub-network by transferring the correct relationships between every two pixels on the feature map. Meanwhile the SRD-loss can further prompt the FDR to recover reliable and accurate fine-grained details and guide the recognition sub-network to perceive the discriminative features from the correct relationships. Extensive experiments on three benchmark datasets and one retail product dataset demonstrate the effectiveness of our proposed framework.

Design of LED grain image recognition and positioning system

LED grain image recognition and location technology is an important technology in LED grain manufacturing and sorting process. Whether the LED grain can be accurately identified and positioned will greatly affect the production speed and efficiency of LED grain. At present, the research on this technology in China is in its infancy. In this paper, a LED grain image recognition and location system based on rectangle detection algorithm is proposed to solve the problem of LED grain image recognition and location. The system uses image grayscale, binarization, edge extraction, contour detection algorithm to process the image of the target image, and uses the matching algorithm and rectangular detection algorithm to identify the LED grain, although the time is short but the accuracy is not enough. Rectangle detection uses two methods to determine contour minima, the getMin() function and the OpenCV library method, using the getMin() function is more appropriate. The LED grain image recognition and positioning system designed in this paper can realize the recognition of LED grain in a simple environment, and can accurately calculate the deflection Angle of LED grain, reducing the time of grain quality detection.

Cluster Representation of the Structural Description of Images for Effective Classification

The problem of image recognition in the computer vision systems is being studied. The results of the development of efficient classification methods, given the figure of processing speed, based on the analysis of the segment representation of the structural description in the form of a set of descriptors are provided. We propose three versions of the classifier according to the following principles: “object–etalon”, “object descriptor–etalon” and “vector description of the object–etalon”, which are not similar in level of integration of researched data analysis. The options for constructing clusters over the whole set of descriptions of the etalon database, separately for each of the etalons, as well as the optimal method to compare sets of segment centers for the etalons and object, are implemented. An experimental rating of the efficiency of the created classifiers in terms of productivity, processing time, and classification quality has been realized of the applied. The proposed methods classify the set of etalons without error. We have formed the inference about the efficiency of classification approaches based on segment centers. The time of image processing according to the developed methods is hundreds of times less than according to the traditional one, without reducing the accuracy.

Explorations of Contrastive Learning in the Field of Small Sample SAR ATR

Synthetic aperture radar (SAR) image recognition plays a significant role in civil and military target perception, and deep learning methods are commonly used for SAR image recognition. However, for small-sample SAR datasets, deep learning methods suffer from problems such as overfitting and gradient explosion during the training process. In this study, first, we pre-train the deep neural network on the unlabeled SAR image dataset with self-supervised contrastive learning, so as to obtain a deep feature extraction network with better representation ability. Then, we transfer the pre-trained weights to the small-sample SAR image data for fine-tuning and linear evaluation, which in turn can improve the overfitting problem of small-sample SAR image recognition. This research can give scientific and objective small-sample target recognition laws, which will promote the long development of the scientific field of deep representation learning research and can generate great economic benefits in the field of SAR image target recognition.

Распознавание микроскопических изображений пыльцевых зерен с помощью сверточной нейронной сети VGG-16

The article presents the result of an experiment on the application of transfer learning using the Visual Geometry Group with 16 layers (VGG-16) convolutional neural network in relation to the problem of recognizing pollen grains in images. An analysis of the information-theoretical base on the application of machine learning algorithms to the problem of classifying pollen grains over the past few years has shown the need to develop (apply) a new method for recognizing images of pollen grains obtained using an optical microscope. Currently, automatic classification for pollen identification is becoming a very active area of research. The article substantiates the task of automating the classification of pollen grains. The aim of the study is to analyze the efficiency and accuracy of classifying microscopic images of pollen grains using transfer learning of the VGG-16 convolutional neural network. Transfer learning was performed using the VGG-16 neural network, which has 13 convolutional layers grouped into 5 blocks with pooling and 3 smoothing layers at the output. Since transfer learning is used, the number of training epochs can be chosen to be small. For this network, only the smoothing output layers change, and the feature extraction remains the same as in the classical model. Therefore, it was chosen to use 10 training epochs. Other hyperparameters are as follows: Drop Out regularization with a probability of 50%, optimization method is ADAM, activation function is sigmoid, loss function is cross-entropy, batch size is 32 images. As a result, by adjusting the hyperparameters of the model and using augmentation, it was possible to achieve a share of correct recognitions of about 80%. At the same time, due to the different number of training examples, the particular characteristics of the classes differ somewhat. Thus, the maximum precision and recall reach 94% and 83%, respectively, for the Dandelion type. In the future, studies are planned to use augmentation as a preprocessing to create a balanced sample. By applying the VGG-16 convolutional neural network to the problem of pollen grain image recognition, high accuracy and efficiency of the method were achieved.

Cell Recognition Using BP Neural Network Edge Computing.

This exploration is to solve the efficiency and accuracy of cell recognition in biological experiments. Neural network technology is applied to the research of cell image recognition. The cell image recognition problem is solved by constructing an image recognition algorithm. First, with an in-depth understanding of computer functions, as a basic intelligent algorithm, the artificial neural network (ANN) is widely used to solve the problem of image recognition. Recently, the backpropagation neural network (BPNN) algorithm has developed into a powerful pattern recognition tool and has been widely used in image edge detection. Then, the structural model of BPNN is introduced in detail. Given the complexity of cell image recognition, an algorithm based on the ANN and BPNN is used to solve this problem. The BPNN algorithm has multiple advantages, such as simple structure, easy hardware implementation, and good learning effect. Next, an image recognition algorithm based on the BPNN is designed and the image recognition process is optimized in combination with edge computing technology to improve the efficiency of algorithm recognition. The experimental results show that compared with the traditional image pattern recognition algorithm, the recognition accuracy of the designed algorithm for cell images is higher than 93.12%, so it has more advantages for processing the cell image algorithm. The results show that the BPNN edge computing can improve the scientific accuracy of cell recognition results, suggesting that edge computing based on the BPNN has a significant practical value for the research and application of cell recognition.

Defending Against Multiple and Unforeseen Adversarial Videos.

Adversarial robustness of deep neural networks has been actively investigated. However, most existing defense approaches are limited to a specific type of adversarial perturbations. Specifically, they often fail to offer resistance to multiple attack types simultaneously, i.e., they lack multi-perturbation robustness. Furthermore, compared to image recognition problems, the adversarial robustness of video recognition models is relatively unexplored. While several studies have proposed how to generate adversarial videos, only a handful of approaches about defense strategies have been published in the literature. In this paper, we propose one of the first defense strategies against multiple types of adversarial videos for video recognition. The proposed method, referred to as MultiBN, performs adversarial training on multiple adversarial video types using multiple independent batch normalization (BN) layers with a learning-based BN selection module. With a multiple BN structure, each BN brach is responsible for learning the distribution of a single perturbation type and thus provides more precise distribution estimations. This mechanism benefits dealing with multiple perturbation types. The BN selection module detects the attack type of an input video and sends it to the corresponding BN branch, making MultiBN fully automatic and allowing end-to-end training. Compared to present adversarial training approaches, the proposed MultiBN exhibits stronger multi-perturbation robustness against different and even unforeseen adversarial video types, ranging from Lp-bounded attacks and physically realizable attacks. This holds true on different datasets and target models. Moreover, we conduct an extensive analysis to study the properties of the multiple BN structure.

Model architecture and tile size selection for convolutional neural network training for non-small cell lung cancer detection on whole slide images

Recent advancements in Artificial-Intelligence-based computer vision systems demonstrated impressive image and pattern recognition capabilities. A special class of neural networks known as Convolutional Neural Networks are used in a wide variety of computer vision tasks such as image classification, object detection and autonomous driving. There is a huge potential to adopt such technology in the domain of pathology. Image data in pathology are considerably larger in size than in typical image recognition problems. Dissecting the image into smaller bits, known as tiling, are often carried out. This paper aims to compare and contrast common model architectures and input tile sizes systematically and find the optimal configuration in the context of lung cancer classification problems.A dataset composed of 87 annotated whole slide images of lung cancer specimens were collected and annotated by two pathologists. Annotated areas were grouped into four classes (Tumor area, Non-tumor area, Necrosis area, and Immune cells). Annotations were converted into labelled tiles at different tile sizes, from 296 to 10000 pixels, (74–2500 μm). The problem was framed as a supervised 4-class classification problem using deep learning.For each tile size, three models, VGG19, InceptionResnetv2 and EfficientNet b3, were trained. Model performances were measured on holdout dataset using standard quantitative metrics including F1-score and AUC-ROC. Our best model instance with tile size at 500 × 500 pixels (125 × 125 μm) achieved an F1-score at 0.9685 and AUC-ROC score at 0.9627.Our results showed that tile size had a significant impact on model performance. The optimal tile size was between 500 and 1000 pixel (125–250 μm) after both quantitative and qualitative assessments. VGG19 marginally outperformed other model architectures.

Glance and Focus Networks for Dynamic Visual Recognition.

Spatial redundancy widely exists in visual recognition tasks, i.e., discriminative features in an image or video frame usually correspond to only a subset of pixels, while the remaining regions are irrelevant to the task at hand. Therefore, static models which process all the pixels with an equal amount of computation result in considerable redundancy in terms of time and space consumption. In this paper, we formulate the image recognition problem as a sequential coarse-to-fine feature learning process, mimicking the human visual system. Specifically, the proposed Glance and Focus Network (GFNet) first extracts a quick global representation of the input image at a low resolution scale, and then strategically attends to a series of salient (small) regions to learn finer features. The sequential process naturally facilitates adaptive inference at test time, as it can be terminated once the model is sufficiently confident about its prediction, avoiding further redundant computation. It is worth noting that the problem of locating discriminant regions in our model is formulated as a reinforcement learning task, thus requiring no additional manual annotations other than classification labels. GFNet is general and flexible as it is compatible with any off-the-shelf backbone models (such as MobileNets, EfficientNets and TSM), which can be conveniently deployed as the feature extractor. Extensive experiments on a variety of image classification and video recognition tasks and with various backbone models demonstrate the remarkable efficiency of our method. For example, it reduces the average latency of the highly efficient MobileNet-V3 on an iPhone XS Max by 1.3x without sacrificing accuracy. Code and pre-trained models are available at https://github.com/blackfeather-wang/GFNet-Pytorch.