Image Recognition Problems Research Articles

Texture Feature Extraction and Recognition of Underwater Target Image Considering Incomplete Tree Wavelet Decomposition

Fu, Y., 2020. Texture feature extraction and recognition of underwater target image considering incomplete tree wavelet decomposition. In: Qiu, Y.; Zhu, H., and Fang, X. (eds.), Current Advancements in Marine and Coastal Research for Technological and Sociological Applications. Journal of Coastal Research, Special Issue No. 107, pp. 25-28. Coconut Creek (Florida), ISSN 0749-0208.Image texture feature extraction has become an important way of image retrieval, scene recognition and target location, which has become an important problem of image extraction and recognition. However, incomplete tree wavelet structure decomposition extraction will get better image clarity, which will be better for processing, such as preprocessing, image noise removal, contrast enhancement and so on. By analyzing the characteristics of image texture, we can decompose the incomplete tree wavelet structure of image, which will better describe the image texture in wavelet domain. In the naval battle, we must find out to launch weapon attack first, which needs the accuracy of automatic recognition of underwater target. Therefore, we need to improve the texture feature extraction and recognition of underwater target image. Firstly, this paper analyzes the image texture features and the algorithm of incomplete tree wavelet decomposition. Then, this paper demonstrates the texture feature extraction and recognition of underwater objects. Finally, some suggestions are put forward.

Sports Video Classification with Deep Convolution Neural Network: A Test on UCF101 Dataset

In the present era, Deep Learning has been applied on a variety of problems from image processing to speech recognition. Convolution Neural Network (CNN) has been extensively used as a powerful classification model for image recognition problems. Video classification presents unique challenges but the problem related to video data is similar to image classification or an object detection problem. The main purpose of video classification in sports is to help the viewers to find the video of their own interest for training and improve the performance. The proposed work is a preliminary attempt to evaluate the performance of deep convolution neural network architectureson the ordered sequence of frames of the sports video. Video classification and video content analysis is one of the ongoing research areas in the field of computer vision. The classification of each frames are recorded and the majority vote of the frames are used to classify the video. UCF101 Video action database has been used for the classification problem.

Model deep learning untuk klasifikasi fragmen metagenom dengan spaced k-mers sebagai ekstraksi fitur

An open challenge in bioinformatics is the analysis of the sequenced metagenomes from the various environments. Several studies demonstrated bacteria classification at the genus level using k-mers as feature extraction where the highest value of k gives better accuracy but it is costly in terms of computational resources and computational time. Spaced k-mers method was used to extract the feature of the sequence using 111 1111 10001 where 1 was a match and 0 was the condition that could be a match or did not match. Currently, deep learning provides the best solutions to many problems in image recognition, speech recognition, and natural language processing. In this research, two different deep learning architectures, namely Deep Neural Network (DNN) and Convolutional Neural Network (CNN), trained to approach the taxonomic classification of metagenome data and spaced k-mers method for feature extraction. The result showed the DNN classifier reached 90.89 % and the CNN classifier reached 88.89 % accuracy at the genus level taxonomy.

Ternary tree-based structural twin support tensor machine for clustering

Most of the real-life applications usually involve complex data, e.g., grayscale images, where information is distributed spatially in the form of two-dimensional matrices (elements of second-order tensor space). Traditional vector-based clustering models such as k-means and support vector clustering rely on low-dimensional features representations for identifying patterns and are prone to loss of useful information which is present in spatial structure of the data. To overcome this limitation, tensor-based clustering models can be utilized for identifying relevant patterns in matrix data as they take advantage of structural information present in multi-dimensional framework and reduce computational overheads as well. However, despite these numerous advantages, tensor clustering has still remained relatively unexplored research area. In this paper, we propose a novel clustering framework, termed as Ternary Tree-based Structural Least Squares Support Tensor Clustering (TT-SLSTWSTC), that builds a cluster model as a hierarchical ternary tree, where at each node non-ambiguous data are dealt separately from ambiguous data points using the proposed Ternary Structural Least Squares Support Tensor Machine (TS-LSTWSTM). The TS-LSTWSTM classifier considers the structural risk minimization of data alongside a symmetrical L2-norm loss function. Also, initialization framework based on tensor k-means has been used in order to overcome the instability disseminated by random initialization. To validate the efficacy of the proposed framework, computational experiments have been performed on human activity recognition and image recognition problems. Experimental results show that our method is not only fast but yields significantly better generalization performance and is comparatively more robust in order to handle heteroscedastic noise and outliers when compared to related methods.

Synthesis of Neural Network Architecture for Recognition of Sea-Going Ship Images

Purpose of research. The current task is to monitor ships using video surveillance cameras installed along the canal. It is important for information communication support for navigation of the Moscow Canal. The main subtask is direct recognition of ships in an image or video. Implementation of a neural network is perspectively.Methods. Various neural network are described. images of ships are an input data for the network. The learning sample uses CIFAR-10 dataset. The network is built and trained by using Keras and TensorFlow machine learning libraries.Results. Implementation of curving artificial neural networks for problems of image recognition is described. Advantages of such architecture when working with images are also described. The selection of Python language for neural network implementation is justified. The main used libraries of machine learning, such as TensorFlow and Keras are described. An experiment has been conducted to train swirl neural networks with different architectures based on Google collaboratoty service. The effectiveness of different architectures was evaluated as a percentage of correct pattern recognition in the test sample. Conclusions have been drawn about parameters influence of screwing neural network on showing its effectiveness.Conclusion. The network with a single curl layer in each cascade showed insufficient results, so three-stage curls with two and three curl layers in each cascade were used. Feature map extension has the greatest impact on the accuracy of image recognition. The increase in cascades' number has less noticeable effect and the increase in the number of screwdriver layers in each cascade does not always have an increase in the accuracy of the neural network. During the study, a three-frame network with two buckling layers in each cascade and 128 feature maps is defined as an optimal architecture of neural network under described conditions. operability checking of architecture's part under consideration on random images of ships confirmed the correctness of optimal architecture choosing.

Classification of breast cancer mammogram images using convolution neural network

SummaryMedical imaging systems have broadly used in the diagnosis and identification of breast cancer. It is essential to recognize breast tumor as soon as possible. Mammography is a widely utilized method for the identification of breast cancer. The identification of cancer is trailed by the segmentation of the cancer area in an image of the mammogram. Numerous researches have been made on the diagnosing and identification of breast cancer utilizing different classification and image processing methods. In this work, we proposed the Convolutional Neural Network (CNN) classifier for diagnosing breast cancer utilizing MIAS (Mammographic Image Analysis Society)‐dataset. CNN established as an efficient class of methods for image recognition problems. CNN is a deep learning system that extricates the feature of an image and utilizes those features for classification of the image. Because deep learning methods are utilized for high task objective Computer Vision, Medical Diagnosis, Image processing, and so on. Wiener filter is utilized to expel the noise and background of the image and the K‐means clustering technique was utilized for the segmentation. After segmentation, the features are extracted and classified utilizing CNN classifier. The performance of this proposed method was analyzed and compared with the conventional techniques based on accuracy, sensitivity, and specificity result parameters. From the comparison result, it is seen that the CNN classifier performed better compared with different techniques with 0.5‐4% additional accuracy and 3‐13% specificity.

The model is a reprint of an object in the image

The problem of facial image recognition (identification) is presented. The difference between facial image recognition and identification is shown. To solve the identification problem, a model of object reprint in the image was developed. This model solves the problem by representing the object in 3-dimensional form, which makes it possible to evaluate and form the necessary characteristics of the object in full, whereas in 2-dimensional form, it is impossible to do this. The model of an object reprint in an image can be used to create a reprint of any spatial objects. To train a reprint model of an object in an image, a multi-layer neural network is used, which is trained sequentially. A local detector for the facial image identification model has been developed to account for acceptable changes in angle, various noise, and different light levels. The binary value that is the result of model processing, represented as activation, determines the relation of a particular image to the corresponding class. The local detector is not only the main element of the model of the object reprint in the image, but it is also a separate mathematical construction. It accepts input data as two-dimensional images. The developed model of object reprint on the image completely solves the problem of identifying a person from the facial image as a whole in conditions of interference and regardless of changes in the angle.

Dimension of the repeller for a piecewise expanding affine map

In this paper, we study the dimension theory of a class of piecewise affine systems in euclidean spaces suggested by Michael Barnsley, with some applications to the fractal image compression. It is a more general version of the class considered in the work of Keane, Simon and Solomyak [The dimension of graph directed attractors with overlaps on the line, with an application to a problem in fractal image recognition. {\it Fund. Math.}, {\bf 180}(3):279-292, 2003] and can be considered as the continuation of the works [On the dimension of self-affine sets and measures with overlaps. {\it Proc. Amer. Math. Soc.}, {\bf 144}(10):4427-4440, 2016], [On the dimension of triangular self-affine sets. {\it Erg. Th. \& Dynam. Sys.}, to appear.] by the authors. We also present some applications of our results for the generalized Takagi functions and fractal interpolation functions.

Removing Noise During the Filtering Images

In the problems of image recognition, various approaches used when the image is noisy and there is a small sample of observations. The article discusses the issue of noise filtering in image processing. The lack of a priori information complicates the processing of data, as a result of which it is necessary to rely on some statistical models of signals and noise. The use of known filters does not always give the desired result. A Gaussian filter can be used for additive noise, a modified Kalman filter eliminates a wider range of noise.

Toward Efficient Image Recognition in Sensor-Based IoT: A Weight Initialization Optimizing Method for CNN Based on RGB Influence Proportion

As the Internet of Things (IoT) is predicted to deal with different problems based on big data, its applications have become increasingly dependent on visual data and deep learning technology, and it is a big challenge to find a suitable method for IoT systems to analyze image data. Traditional deep learning methods have never explicitly taken the color differences of data into account, but from the experience of human vision, colors play differently significant roles in recognizing things. This paper proposes a weight initialization method for deep learning in image recognition problems based on RGB influence proportion, aiming to improve the training process of the learning algorithms. In this paper, we try to extract the RGB proportion and utilize it in the weight initialization process. We conduct several experiments on different datasets to evaluate the effectiveness of our proposal, and it is proven to be effective on small datasets. In addition, as for the access to the RGB influence proportion, we also provide an expedient approach to get the early proportion for the following usage. We assume that the proposed method can be used for IoT sensors to securely analyze complex data in the future.

A self-training hierarchical prototype-based approach for semi-supervised classification

This paper introduces a novel self-training hierarchical prototype-based approach for semi-supervised classification. The proposed approach firstly identifies meaningful prototypes from labelled samples at multiple levels of granularity and, then, self-organizes a highly transparent, multi-layered recognition model by arranging them in a form of pyramidal hierarchies. After this, the learning model continues to self-evolve its structure and self-expand its knowledge base to incorporate new patterns recognized from unlabelled samples by exploiting the pseudo-label technique. Thanks to its prototype-based nature, the overall computational process of the proposed approach is highly explainable and traceable. Experimental studies with various benchmark image recognition problems demonstrate the state-of-the-art performance of the proposed approach, showing its strong capability to mine key information from unlabelled data for classification.

The pattern recognition problem solution using a convolutional neural network

The article deals with the issues of theoretical and applied modeling based on convolutional neural network in the field of agricultural production. In particular, the specificity of convolutional neural network is considered, its model, architecture and layers are presented. Possibilities of its use for the solution of problems of recognition and classification of images are revealed. In addition, it was found that the convolutional neural network differs from the usual perceptron in that each neural layer is not connected to all the neurons of the previous layer, but only to a part of it. In networks of this type, a small matrix is used for the convolution operation, which moves along the processed layer with a certain offset step. For the first layer, this matrix moves across the input image. After each shift, a signal is generated to activate the neuron of the next layer and from the corresponding position. As a result of studying the problems of image recognition using electronic computers, the possibilities of machine learning as a method of recognizing an object in an image that characterizes each object with a set of features are revealed. In addition, the author reveals the specifics of using the convolutional neural network mathematical apparatus as a model for image recognition, which consists in using an array of data describing the image and applying the method of error back propagation.

A Stacking-based Ensemble Learning Method for Outlier Detection

Outlier detection is considered as one of the crucial research areas for data mining. Many methods have been studied widely and utilized for achieving better results in outlier detection from existing literature; however, the effects of these few ways are inadequate. In this paper, a stacking-based ensemble classifier has been proposed along with four base learners (namely, Rotation Forest, Random Forest, Bagging and Boosting) and a Meta-learner (namely, Logistic Regression) to progress the outlier detection performance. The proposed mechanism is evaluated on five datasets from the ODDS library by adopting five performance criteria. The experimental outcomes demonstrate that the proposed method outperforms than the conventional ensemble approaches concerning the accuracy, AUC (Area Under Curve), precision, recall and F-measure values. This method can be used for image recognition and machine learning problems, such as binary classification.

Face image recognition based on convolutional neural network

With the continuous progress of The Times and the development of technology, the rise of network social media has also brought the explosive growth of image data. As one of the main ways of People's Daily communication, image is widely used as a carrier of communication because of its rich content, intuitive and other advantages. Image recognition based on convolution neural network is the first application in the field of image recognition. A series of algorithm operations such as image eigenvalue extraction, recognition and convolution are used to identify and analyze different images. The rapid development of artificial intelligence makes machine learning more and more important in its research field. Use algorithms to learn each piece of data and predict the outcome. This has become an important key to open the door of artificial intelligence. In machine vision, image recognition is the foundation, but how to associate the low-level information in the image with the high-level image semantics becomes the key problem of image recognition. Predecessors have provided many model algorithms, which have laid a solid foundation for the development of artificial intelligence and image recognition. The multi-level information fusion model based on the VGG16 model is an improvement on the fully connected neural network. Different from full connection network, convolutional neural network does not use full connection method in each layer of neurons of neural network, but USES some nodes for connection. Although this method reduces the computation time, due to the fact that the convolutional neural network model will lose some useful feature information in the process of propagation and calculation, this paper improves the model to be a multi-level information fusion of the convolution calculation method, and further recovers the discarded feature information, so as to improve the recognition rate of the image. VGG divides the network into five groups (mimicking the five layers of AlexNet), yet it USES 3∗3 filters and combines them as a convolution sequence. Network deeper DCNN, channel number is bigger. The recognition rate of the model was verified by 0RL Face Database, BioID Face Database and CASIA Face Image Database.

Aeroengine Control System Sensor Fault Diagnosis Based on CWT and CNN

The aeroengine control system is a piece of complex thermal machinery which works under high-speed, high-load, and high-temperature environmental conditions over lengthy periods of time; it must be designed for the utmost reliability and safety to function effectively. The consequences of sensor faults are often extremely serious. The inherent complexity of the engine structure creates difficulty in establishing accurate mathematical models for the model-based sensor fault diagnosis. This paper proposes an intelligent fault diagnosis method for aeroengine sensors combining a deep learning algorithm with time-frequency analysis wherein the signal recognition problem is transformed into an image recognition problem. The continuous wavelet transform (CWT) is first applied to seven common health condition signals in an engine control system sensor in order to generate scalograms that capture the characteristics of the signal. A convolutional neural network (CNN) model trained with preprocessed and labeled datasets is then used to extract the features of a time-frequency graph based on which faults can be identified and isolated. This method does not require modeling and design thresholds, so it has strong robustness and accuracy rate of over 97%. The trained model effectively reveals faults in sensor signals and allows for accurate identification of fault types.

Numbering and Classification of Panoramic Dental Images Using 6-Layer Convolutional Neural Network

Deep Convolution Neural Network is one of the most powerful tools to solve complex problems of image classification, image recognition, financial analysis, medical diagnosis and many similar problems. A dental panoramic image consists of collection of teeth of both upper jaw and lower jaw. Automatic classification of dental panoramic images into various tooth types such as canines, incisors, premolars and molars has been a challenging task and involves crucial role of an experienced dentist. In this paper, we propose a technique for numbering and classification of the panoramic dental images. The proposed algorithm consists of four stages namely pre-processing, segmentation, numbering and classification. The pre-processed panoramic dental images are segmented using fuzzy c-mean clustering and subjected to vertical integral projection to extract a single tooth. The image dataset consists of 400 dental panoramic images collected from various dental clinics. The 400 dental images are divided into 240 training samples and 160 testing samples. The image data set is augmented by applying various transformations. Panoramic dental images are further numbered using a universal dental numbering system. Finally, the classification is done with the help of 6-layer deep convolution neural network (DCNN) consisting of 3 convolutional neural network and 3 fully connected network. The tooth is classified as canine, incisor, molar and premolar. An accuracy of 95% has been achieved for augmented database and 92% for original dataset with the proposed algorithm. The proposed numbering and classification of dental panoramic images is useful in biomedical application and postmortem recording of dental records. In case of big calamity, the system can also assist the dentist in recording post mortem dental record that is a very lengthy and arduous task.

ЗАСТОСУВАННЯ ЗГОРТКОВИХ НЕЙРОННИХ МЕРЕЖ ДЛЯ БЕЗПЕКИ РОЗПІЗНАВАННЯ ОБ'ЄКТІВ У ВІДЕОПОТОЦ

The article is devoted to analyzing methods for recognizing images and finding them in the video stream. The evolution of the structure of convolutional neural networks used in the field of computer video flow diagnostics is analyzed. The performance of video flow diagnostics algorithms and car license plate recognition has been evaluated. The technique of recognizing the license plates of cars in the video stream of transport neural networks is described. The study focuses on the creation of a combined system that combines artificial intelligence and computer vision based on fuzzy logic. To solve the problem of license plate image recognition in the video stream of the transport system, a method of image recognition in a continuous video stream with its implementation based on the composition of traditional image processing methods and neural networks with convolutional and periodic layers is proposed. The structure and peculiarities of functioning of the intelligent distributed system of urban transport safety, which feature is the use of mobile devices connected to a single network, are described. A practical implementation of a software application for recognizing car license plates by mobile devices on the Android operating system platform has been proposed and implemented. Various real-time vehicle license plate recognition scenarios have been developed and stored in a database for further analysis and use. The proposed application uses two different specialized neural networks: one for detecting objects in the video stream, the other for recognizing text from the selected image. Testing and analysis of software applications on the Android operating system platform for license plate recognition in real time confirmed the functionality of the proposed mathematical software and can be used to securely analyze the license plates of cars in the scanned video stream by comparing with license plates in the existing database. The authors have implemented the operation of the method of convolutional neural networks detection and recognition of license plates, personnel and critical situations in the video stream from cameras of mobile devices in real time. The possibility of its application in the field of safe identification of car license plates has been demonstrated.

FOSNet: An End-to-End Trainable Deep Neural Network for Scene Recognition

Scene recognition is an image recognition problem aimed at predicting the category of the place at which the image is taken. In this paper, a new scene recognition method using the convolutional neural network (CNN) is proposed. The proposed method is based on the fusion of the object and the scene information in the given image and the CNN framework is named as FOS (fusion of object and scene) Net. In addition, a new loss named scene coherence loss (SCL) is developed to train the FOSNet and to improve the scene recognition performance. The proposed SCL is based on the unique traits of the scene that the 'sceneness' spreads and the scene class does not change all over the image. The proposed FOSNet was experimented with three most popular scene recognition datasets, and their state-of-the-art performance is obtained in two sets: 60.14% on Places 2 and 90.37% on MIT indoor 67. The second highest performance of 77.28% is obtained on SUN 397.

DETECTION OF LEUKEMIA IN MICROSCOPIC WHITE BLOOD CELL IMAGES USING GAUSSIAN FEATURE CONVOLUTIONAL VISUAL RECOGNITION ALGORITHM

Cancer detection has always been a challenge in the diagnosis and treatment plan for hematological diseases. Manual identification of cancers from microscopic images is necessary. It depends on other factors, including the lack of microscopic images as their knowledge and precise measurement of specific usual or cancer expert classifications. Automated identification of cancer cells from microscopic section images and helps to improve the problem mentioned above, based on biological interpretation. Detection of blood cancer cells by a patient's blood mark microscopic examination using different techniques of image processing. White blood cell cancer images and their essential three-stage there are an enhancement, segmentation, and classification. The stage of image processing to achieve more quality and accuracy in detecting blood cancers. Approaching the conventional segmentation and counting of blood cells is considered as an essential step that helps to extract features to diagnose diseases like leukemia. The image analysis will allow hematologist experts to perform faster and more accurately. In this work, the image recognition problem of White Blood Cells (WBC) cancer is investigated. Different types of white blood cells are classified using a Gaussian Feature Convolutional Visual Recognition (GFCVR).The most important features or segments of these blood cells are provided as input microscopic images to the neural network. The consequences of this analysis are mainly considering the identification of blood cancer affected region with feature extraction for specific white blood cell images. In this proposed approach pre-processing the image, classifies multi-level clustering to find the affected area in blood cells. In this proposed simulation, results in time complexity, accuracy, recall, perception, false classification ratio gives better results compare with other existing methods.

Capsule network with K-Means routingfor plant disease recognition

Capsule Networks (CapsNets) excel on simple image recognition problems. However, they fail to perform on complex images with high similarity and background objects. This paper proposes Local Binary Pattern (LBP) k-means routing and evaluates its performance on three publicly available plant disease datasets containing images with high similarity and background objects. The proposed routing algorithm adopts the squared Euclidean distance, sigmoid function, and a ‘simple-squash’ in place of dot product, SoftMax normalizer, and the squashing function found respectively in the dynamic routing algorithm. Extensive experiments conducted on the three datasets showed that the proposed model achieves consistent improvement in test accuracy across the three datasets as well as allowing an increase in the number of routing iterations with no performance degradation. The proposed model outperformed a baseline CapsNet by 8.37% on the tomato dataset with an overall test accuracy of 98.80%, comparable to state-of-the-art models on the same datasets.