Image Recognition Algorithm Research Articles

Hierarchical Traffic Sign Recognition System

A hierarchical system of classifiers for recognizing traffic signs from their images is proposed in the article. Only signs that fit into a square are considered. The traffic signs set is analyzed by their shape, color and basic features. A hierarchy of classes for traffic signs on the details of their images is proposed. The traffic sign image recognition algorithm uses this hierarchy. The algorithm only works with localized signs. The localization algorithm is not considered. Image preprocessing is performed at each level of the hierarchy for traffic sign features. Different classifiers are used at different levels of the hierarchy. Preprocessing at the first level uses the segmentation method. Signs are classified by their form in the first level. The edge segment is highlighted in the image of the traffic sign. The Euclidean distance in color space is used to calculate the edge segment. The edge segment outline defines the shape of the sign. Classification of the sign by its shape is carried out based on comparison with the standards. The classifier defines the pattern closest to the sign contour. The efficiency of the classification in the first level is 99.8 %. The second level classifies signs by the color of the edge segment. The effectiveness of this classification is close to 100 %. Internal sign images are obtained after the edge segment is removed from the image. The following levels classify by internal images. These classifiers use artificial neural networks. The efficiency of the system on the GTSRB collection is 98 %.

Image Recognition and Encryption Algorithm Based on Artificial Neural Network and Multidimensional Chaotic Sequence.

With the continuous development of Internet technology and technological innovation, image recognition technologies such as face unlocking and face brushing payment have gradually entered daily life. However, it can not be ignored that these technologies not only bring us great convenience but also face great risks. The biological characteristics of a face image are unique, and it will be difficult to modify once it is leaked. If the image information stored in the cloud is leaked because it cannot be properly kept, users have no privacy. The encryption and recognition of face image can effectively solve this problem. Aiming at this, high-dimensional chaos Henon Map and one-dimensional chaos Logistic Map are used to generate a key to complete the encryption of the image in the transformation domain, and the capacity and complexity of the key are further enhanced. Then, combined with BP neural network to achieve face image recognition. Finally, the robustness of the proposed algorithm is verified and analyzed by conventional attacks, geometric attacks, and occlusion attacks.

TOURISM, NATURAL PROTECTED AREAS AND OPEN SOURCE GEOSPATIAL TECHNOLOGIES

Abstract. The Covid-19 outbreak has greatly impacted society behaviours fostering proximity tourism and valorising the social role of peri-urban natural protected areas as key locations for outdoor activities. FOSS and FOSS4G can play a critical role to support the value creation for these sites. This work evaluates its application in the context of two different protected areas for the creation of 3D digital products, the monitoring of touristic fluxes and the conduction of parks management activities. To this aim three solutions that copes with the mentioned aspects are presented and gaps, weakness and limitations evaluated. The investigated solutions consists in: the data workflow from survey to 3D rendering using Blender and GIS plugin; the touristic fluxes monitoring system based on a machine learning algorithm for image recognition from captured video data streams and istSOS; and finally the park assets management system which is based on PostGIS and OpenLayers.

EXTERNAL CRITERION OF STABILITY OF INTRA-MULTIPLE DISTANCES IN TASKS OF DIAGNOSIS OF STATES OF TECHNICAL OBJECTS

The purpose of the research is adaptation of biclasticization criteria, which have the properties of external complement to the problems of classification of states of complex technical objects in the field of computer engineering. The principles of inductive modelling of complex systems, in particular, the principle of external complementarity, as well as the methodology of the theory of pattern recognition, methods of inductive cluster analysis, mathematical statistics are applied. The construction of the criterion of stability of intra-multiple distances is based on the schemes of known criteria of self-organization of models, in particular the criterion of model inconsistencies, which are widely used in multi-step image recognition algorithms with intelligent choice of optimal results. The paper proposes the application of the criterion of stability of intra-multiple distances in the problems of diagnosing the state of technical objects, in particular, in the field of computer engineering. Since the application of such a criterion requires the presence of a target feature, one of the options for splitting the original experimental database into two subsets has been adapted: a subset of target features and a subset of input parameters. The concept of intra-multiple distances is extended to the application of algorithms of optimal complexity in the criteria of self-organization algorithms. Intelligent algorithms for self-organization of models of optimal complexity can be used to increase the reliability of computer systems.

Machine Learning and Image Recognition Technologies to Identify Built Environment Barriers and Incentives to Walk

The local built environment characteristics deeply influence pedestrians’ behavior, favoring or imposing barriers on walking trips. However, identifying micro-scale built environment data is challenging and time consuming, and in developing countries there is a general lack of reliable information at street level. The recent development of machine learning and image recognition algorithms is helping researchers to collect data quickly, automatically, and on a large scale. Therefore, this study aims to test the application of an existing semantic segmentation algorithm to represent urban scenes in the city of São Paulo. A confirmatory multivariate technique (structural equation model [SEM]) is used to test if a combination of the predetermined categories derived from machine learning algorithms helps to understand which type of environment (urban scenes) represents barriers or incentives to walk. The impacts of the urban scenes on the walking behavior mediated by the walkability perceptions were tested using the aforementioned SEM model. Car-oriented and unoccupied areas, with a high presence of heavy vehicles and a large presence of vegetation, are considered detrimental to the walkability perception and consequently to the walking frequency. On the other hand, densified areas, proximity to public transportation routes, and the presence of lighting and other pedestrians are considered friendlier for pedestrians, encouraging residents to walk.

Study on Morphological Identification of Tight Oil Reservoir Residual Oil after Water Flooding in Secondary Oil Layers Based on Convolution Neural Network

In this paper, a microscopic oil displacement visualization experiment based on the glass etching model to simulate the tight oil reservoir of underground rocks is carried out. At present, water flooding technology is widely used in the development of oil and gas fields, and the remaining oil content is still very high after water flooding. It is the key to improving oil recovery to identify and study the remaining oil form distribution after water flooding. The experiment result shows there are five types of residual oil after water flooding: columnar residual oil, membranous residual oil, oil droplet residual oil, blind terminal residual oil and cluster residual oil. A convolution neural network is suitable for complex image characteristics with good robustness. In recent years, it has made a breakthrough in a set of small and efficient neural networks with SqueezeNet, Google Inception and the flattened network method put forward. In order to solve the problems of low automation, low efficiency and high error rate in the traditional remaining oil form recognition algorithm after water flooding in tight oil reservoirs, an image recognition algorithm based on the MobileNets convolutional neural network model was proposed in this paper to achieve accurate recognition of the remaining oil form. Based on traditional image processing methods which, respectively, extracted the whole picture of the different types of remaining oil in the image block, it uses the MobileNets network structure to classify different types of image block and realizes the layered depth convolution neural network system. The experiment result shows that the model can accurately identify the remaining oil forms, and the overall recognition accuracy is up to 83.8% after the convergence of the network model, which infinitely identifies the remaining oil forms in the morphological library, proving the strong generalization and robustness of the model.

Multimedia sensor image detection based on constrained underdetermined equation

Abstract In order to study the image detection of multimedia sensor based on constrained under determined equation, this paper proposes the improvement of image sensor quality detection method based on image recognition algorithm. Firstly, the principle and steps of the under determined method are explained, and the complexity of the algorithm is analyzed. The image sensor is the most important part of the mobile camera, which determines the performance of the mobile camera, and its assembly quality largely determines the quality of the whole mobile camera. The case studied in this paper is model h mobile phone camera, and the first pass rate of good products is only 80%. In order to meet the 95% pass rate of good products required by customers, it is necessary to comprehensively improve the processes that have a great impact on product quality, so as to improve the first pass rate of products, achieve quality objectives and reduce manufacturing failure costs. In order to find out the threshold value of the height information measured by the improved method, to determine whether the product is good, and to find out the process that has a great impact on the surface flatness of the image sensor, the product measurement experiment is specially set. The improved war page test method was used to measure the products of each experimental group. Among them, No. 22, No. 39 and No. 40 products are poor due to changes in physical properties in subsequent production, and cannot flow to the next process for production. Therefore, the data of No. 22, No. 39 and No. 40 are missing in the test data. A total of 133 groups of experimental data were obtained in this experimental test. The results showed that the standard deviations of group 1, group 2, group 3 and group 4 were 0.172, 0.125 and 0.304 respectively. That is, after the same product is treated by false hardening, heating hardening and base installation processes, the false hardening and heating hardening processes have relatively little impact on the surface flatness of the image sensor. After the base opening and installation process, the surface flatness of the image sensor begins to change. The standard deviations of the measurement results of group 1 and groups 4, 5, 6 and 7 were 0.304, 0.381, 0.391 and 0.514 respectively. That is, after the base installation, calibration, bonding, soldering and testing of the product, the surface flatness of the image sensor and the initial surface flatness have changed greatly. Therefore, production technicians can focus on the process that starts to change, that is, the base installation process. The improved method has the advantages of automatic operation, low error, low error rate and so on.

Image Target Detection and Recognition Method Using Deep Learning

Image target detection and recognition had been widely used in many fields. However, the existing methods had poor robustness; they not only had high error rate of target recognition but also had high dependence on parameters, so they were limited in application. Therefore, this paper proposed an image target detection and recognition method based on the improved R-CNN model, so as to detect and recognize the dynamic image target in real time. Based on the analysis of the existing theories of deep learning detection and recognition, this paper summarized the composition and working principle of the traditional image target detection and recognition system and compared the basic models of target detection and recognition, such as R-CNN network, Fast-RCNN network, and Faster-RCNN network. In order to improve the accuracy and real-time performance of the model in image target detection and recognition, this paper adopted the target feature matching module in the existing R-CNN network model, so as to obtain the feature map close to the same target through similarity calculation for the features extracted by the model. Therefore, an image target detection and recognition algorithm based on the improved R-CNN network model is proposed. Finally, the experimental results showed that the image target detection and recognition algorithm proposed in this paper can be better applied to image target detection and classification in complex environment and had higher detection efficiency and recognition accuracy than the existing models. The target detection and recognition algorithm proposed in this paper had certain reference value and guiding significance for further application research in related fields.

Image Recognition Algorithm Based on Artificial Intelligence and Machine Learning

At present, most of the existing image recognition methods are not only cumbersome in process but also require manual design of functions, resulting in poor recognition results and time-consuming training. This study explores image recognition algorithms based on artificial intelligence and machine learning, which can simulate the hierarchical structure of the human brain and nervous system, realize automatic extraction of complex features, and have powerful data representation capabilities. In this study, the structure of artificial neuron is introduced first, and the training process of the neural network is introduced in detail. At the same time, this study improves the traditional training algorithm and proposes two new machine learning models for different application scenarios, which effectively improves the performance of the optimal model. In relevant tests, the effect is significant. On the basis of these improved algorithms, an online recognition system is designed and implemented. The recognition accuracy rate of the system for the hidden layer reaches more than 90%, which verifies the effectiveness of the technology. At the same time, after repeated experiments, the results show that the multiview algorithm effectively solves the problem that the recognition results of the traditional multiview algorithm are affected by the size of the target contour.

Application of Intelligent Technology in Facade Style Recognition of Harbin Modern Architecture

The judgment of facade styles is an important part of the sustainable utilization and restorative process of historical architectures. Contemporary Harbin needs the help of modern architectural facade forms in the planning of the famous historic city, especially with the facade renovation of old architectures with non-cultural heritage. This paper discusses the possibility of applying advanced image recognition algorithms to the classification of the modern Harbin architectural facade styles and argues that the keys to the classification and positioning of the styles are the forms, the details, and the decorative patterns of the architectural facades, together with the deformation and the quantitative variation factors of the facade decoration symbols. Based on the conventional classification method, the facade styles of Harbin modern architecture were divided into 12 categories after data analysis. To better capture the overall structure information and the style features of the local components in the architectural images, the group convolution and the dilated convolution were added into the ResNet model, and then, the improved channel attention mechanism was introduced to construct a novel CA-MSResNet model. The CA-MSResNet model could more accurately identify the morphological elements and the style categories of the architectures, and the average accuracy reached 87.5%. These techniques, with their promising results, are expected to be further applied in the future research on the sustainable utilization and renovation of Harbin modern architecture.

Dermoscopic Image Classification of Pigmented Nevus under Deep Learning and the Correlation with Pathological Features.

The objective of this study was to explore the image classification and case characteristics of pigmented nevus (PN) diagnosed by dermoscopy under deep learning. 268 patients were included as the research objects and they were randomly divided into observation group (n = 134) and control group (n = 134). Image recognition algorithm was used for feature extraction, segmentation, and classification of dermoscopic images, and the image recognition and classification algorithm were studied as the performance and accuracy of fusion classifier were compared. The results showed that the classifier was optimized, and the linear kernel accuracy was 85.82%. The PN studied mainly included mixed nevus, junctional nevus, intradermal nevus, and acral nevus. The sensitivity under collaborative training was higher than that under feature training and fusion feature training, and the differences among three trainings were significant (P < 0.05). The sensitivity of the observation group was 88.65%, and the specificity was 90.26%, while the sensitivity and the specificity of the control group were 85.65% and 84.03%, respectively; there were significant differences between the two groups (P < 0.05). In conclusion, dermoscopy under deep learning could be applied as a diagnostic way of PN, which helped improve the accuracy of diagnosis. The dermoscopic manifestations of PN showed a certain corresponding relationship with the type of cases and could provide auxiliary diagnosis in clinical practice. It could be applied clinically.

Complex Network Study of Solar Magnetograms.

In this paper, we study solar magnetic activity by means of a complex network approach. A complex network was built based on information on the space and time evolution of sunspots provided by image recognition algorithms on solar magnetograms taken during the complete 23rd solar cycle. Both directed and undirected networks were built, and various measures such as degree distributions, clustering coefficient, average shortest path, various centrality measures, and Gini coefficients calculated for all them. We find that certain measures are correlated with solar activity and others are anticorrelated, while several measures are essentially constant along the solar cycle. Thus, we show that complex network analysis can yield useful information on the evolution of solar activity and reveal universal features valid at any stage of the solar cycle; the implications of this research for the prediction of solar maxima are discussed as well.

Image Processing and Recognition Algorithm Design in Intelligent Imaging Device System

People may quickly employ imagig devices to acquire and use image data thanks to the rapid development of computer networks and communication technologies. However, imaging devices obtain massive data through real-time acquisition, and a large number of invalid images affect the imaging device system’s endurance on the one hand while also requiring a significant amount of time for analysis on the other hand, so there is a critical need to find a way to automate the mining of valuable information in the data. In this paper, we propose an intelligent imaging device system, which embeds a target intelligent recognition algorithm, improves the YOLOv3 model by using a method based on depth-separable convolutional blocks and inverse feature fusion structure, and finally achieves fast target detection while improving detection accuracy through the design of distance-based nonextreme suppression and loss function. By preprocessing the images and automatically identifying and saving images containing target animals, the range of the imaging device system equipment can be improved and the workload of researchers searching for target animals in images can be reduced. In this paper, we propose a method for intelligent preservation of contained target images by deploying lightweight target recognition algorithms on edge computing hardware with the intelligence of the intelligent imaging device system as the research goal. After simulation experiments, the use of this method can improve the endurance of the imaging device system and reduce the time of manual processing at a later stage.

E-Tools for Personalizing Learning During the Pandemic: Case Study of an Innovative Solution for Remote Teaching.

“Every child counts” has lost its value even from the political discourse of some societies during the pandemic, proving that the level of culture of inclusion is the true standard of Sustainable Development Goal (SDG) commitment. Online education and therapy required rethinking the way we educate children with special needs and, implicitly, prepare them for life. We consider that the personalized approach of the therapeutic intervention was the main difficulty. In this article, we propose a solution to this problem, an approach based on a platform initially developed by tactileimages.org for vision-impaired pupils which became a tool in the universal design of learning materials. This e-learning tool includes an Editor, a browser-based software developed to allow the creation or adaptation of drawings into vector images; the QR code through which areas of educational and therapeutic interest are allocated to pictures for task personalization; and the voice-over function of the companion application. The customized material is identified by image recognition algorithms, and the user's gesture is recognized by artificial intelligence algorithms, which receives (by voice-over) details about therapeutic tasks in remote teaching. The article illustrates the personalization of the therapeutic and educational path. The process starts with defining the child's functioning profile and matching function with the curriculum elements as they are found within the Erasmus + project “Cognitive Resources for Toddlers Teens and Experts” —stored in the virtual library. Information and comunication technology is currently an important vector in attaining the SDG vision. The proposed solution will be improved in order to further personalize the educational and therapeutic intervention in the post-pandemic period too.

Tell Me More: Automating Emojis Classification for Better Accessibility and Emotional Context Recognition

Users of web or chat social networks typically use emojis (e.g., smilies, memes, hearts) to convey in their textual interactions the emotions underlying the context of the communication, aiming for better interpretability, especially for short polysemous phrases. Semantic-based context recognition tools, employed in any chat or social network, can directly comprehend text-based emoticons (i.e., emojis created from a combination of symbols and characters) and translate them into audio information (e.g., text-to-speech readers for individuals with vision impairment). On the other hand, for a comprehensive understanding of the semantic context, image-based emojis require image-recognition algorithms. This study aims to explore and compare different classification methods for pictograms, applied to emojis collected from Internet sources. Each emoji is labeled according to the basic Ekman model of six emotional states. The first step involves extraction of emoji features through convolutional neural networks, which are then used to train conventional supervised machine learning classifiers for purposes of comparison. The second experimental step broadens the comparison to deep learning networks. The results reveal that both the conventional and deep learning classification approaches accomplish the goal effectively, with deep transfer learning exhibiting a highly satisfactory performance, as expected.

Star Identification Based on Radial Triangle Mapping Matrix

In this paper, a novel star image recognition algorithm is proposed based on a voting scheme that utilizes a radial triangle mapping matrix pattern. Inspired by the radial and cyclic algorithm, the radial triangle mapping matrix pattern is rotation-invariant, and does not rely on a location star, which provides a new idea for the construction of rotation-invariant patterns. The algorithm is based on a geometric voting scheme in which the radial triangle group is considered as the basic voting element. The database optimization is used to reduce the memory requirement and improve the query efficiency by extracting the nonzero value of the pattern. During the identification period, the voting scheme can vote for the candidate stars through the radial triangles, which improves the robustness towards the noise. The proposed algorithm exhibits a 98.41&#x0025; identification rate with 2.0-pixel standard deviation of position noises and 0.3-Mv magnitude noises. The experimental results of the simulation star image and real star image show that the proposed algorithm is quite more robust towards position noise, magnitude noise, and false star noise when compared with other comparing algorithms.

Drone-Aided Path Planning for Unmanned Ground Vehicle Rapid Traversing Obstacle Area

Even with visual contact equipment such as cameras, an unmanned ground vehicle (UGV) alone usually takes a lot of time to navigate an unfamiliar area with obstacles. Therefore, this study proposes a fast drone-aided path planning approach to help UGVs traverse an unfamiliar area with obstacles. In this scenario called UAV/UGV mobile collaboration (abbreviated UAGVMC), a UGV initially invokes an unmanned aerial vehicle (UAV) at the scene to take a ground image and send it back to the cloud to proceed with object detection, image recognition, and path planning (abbreviated odirpp). The cloud then sends the UGV a well-planned path map to help traverse an unfamiliar area. This approach uses the one-stage object detection and image recognition algorithm YOLOv4-CSP to quickly and accurately identify obstacles and the New Bidirectional A* (NBA*) algorithm to plan an optimal route avoiding ground objects. Experiments show that the execution time of path planning for each scene is less than 10 s on average. It does not affect the image quality of the path map. It ensures that the user can correctly interpret the path map and remotely drive the UGV rapidly, passing through that unfamiliar area with obstacles. As a result, the selected model can outperform the other alternatives significantly by average performance ratio up to 3.87 times on average.

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%.

The Influence of Intelligent Visual Sensing Technology on Online English Teaching in Wireless Network Environment

The study is aimed at improving the quality of online English teaching and enhancing students’ English learning ability and effect. First, the current situation of online English teaching is analyzed. Second, the image recognition algorithm in artificial intelligence (AI) visual sensing technology is used to process the image, and the effects of illumination and image acquisition angle are analyzed. Finally, AI image recognition technology is applied to online English teaching, and its impact on online English teaching is analyzed. The results show that AI visual sensing technology can accurately recognize students’ facial information. Therefore, it can confirm the identity of students, help to count the number of students in online English teaching, and improve students’ autonomous learning ability. In addition, it can also record students’ facial information in real time and analyze students’ emotional changes in the learning process to adjust the teaching process and improve the teaching quality in an all-round way. The results provide technical support for improving online English teaching and contribute to English teaching reform.

A New Method of Conductor Galloping Monitoring Using the Target Detection of Infrared Source

Because the galloping of iced conductors is one of the main disasters in the State Grid, resulting in huge economic and property losses every year, the research on relevant monitoring methods is of great significance. The existing galloping monitoring technology is mainly based on the contact detection method, which presents potential electrical hazards and power supply problems. In this paper, a conductor galloping monitoring method based on the target detection of infrared sources is put forward to overcome the shortcomings of existing methods. In other words, an infrared source label is installed on the conductor spacer, high-definition night vision infrared cameras are installed on electric power towers to take video of the infrared source labels, and the characteristic amplitude of conductor galloping is calculated by an image recognition and tracking algorithm. The practical application results indicate that the method has the advantages of non-contact detection, safety and reliability, and high detection accuracy.