Object Recognition Method Research Articles

Road object recognition method based on improved YOLOv3

Based on the emergence and development of autonomous driving technology, the identification of obstacles on the road is a very important and challenging task. And there are many difficulties in the realization of this task, for example, there are many types of road targets, and the scale span of the road obstacles is large. In view of these problems, this experiment proposes three improvement directions for the YOLOv3 algorithm to perform the task of road target prediction: one is to improve the up-sampling multiple to use more shallow spatial information to improve the accuracy of small target detection. The second is to change the way of feature fusion of the feature pyramid. Thirdly, the convergence direction of the model is changed by clustering learning. Experiments on the BDD100K data set show that the yolov3_10cls_tiny algorithm proposed in this paper has the best detection performance and better average accuracy than YOLOv3.

Real-Time Automatic Investigation of Indian Roadway Animals by 3D Reconstruction Detection Using Deep Learning for R-3D-YOLOv3 Image Classification and Filtering

Statistical reports say that, from 2011 to 2021, more than 11,915 stray animals, such as cats, dogs, goats, cows, etc., and wild animals were wounded in road accidents. Most of the accidents occurred due to negligence and doziness of drivers. These issues can be handled brilliantly using stray and wild animals-vehicle interaction and the pedestrians’ awareness. This paper briefs a detailed forum on GPU-based embedded systems and ODT real-time applications. ML trains machines to recognize images more accurately than humans. This provides a unique and real-time solution using deep-learning real 3D motion-based YOLOv3 (DL-R-3D-YOLOv3) ODT of images on mobility. Besides, it discovers methods for multiple views of flexible objects using 3D reconstruction, especially for stray and wild animals. Computer vision-based IoT devices are also besieged by this DL-R-3D-YOLOv3 model. It seeks solutions by forecasting image filters to find object properties and semantics for object recognition methods leading to closed-loop ODT.

Image and model sequences matching for on-site assembly stage identification

In an assembly task, stage identification determines the subsequent parts to be assembled and therefore, it affects the quality and efficiency of the entire assembly operation in computer-aided manufacturing. However, due to the similarity between the assemblies in two adjacent stages, it is difficult to distinguish the assembly accurately based on the current visual recognition method. Considering the time and progress logic of assembly procedures, a stage identification method based on the on-site assembly sequence and stage model sequence is proposed. The on-site assembly sequence is generated with the assembly progress to record the history information in the time order and the stage model sequence is preprocessed in the database in the progress order. On the one hand, these two sequences provide the constraints to avoid the inefficient and indistinguishable assembly images to be analyzed. On the other hand, the two sequences contain time and assembly progress logic, which can be used to solve the stage identification issue using an optimal mapping model, so as to improve the identification efficiency. This paper introduces the methodology from three aspects, namely, offline stage model sequence generation, online assembly sequence matching and assembly stage identification. The experiments proved the proposed method can effectively improve the accuracy of real-time assembly stage identification compared with the original object recognition method.

Joint Feature-Space and Sample-Space Based Heterogeneous Feature Transfer Method for Object Recognition Using Remote Sensing Images with Different Spatial Resolutions.

To improve the classification results of high-resolution remote sensing images (RSIs), it is necessary to use feature transfer methods to mine the relevant information between high-resolution RSIs and low-resolution RSIs to train the classifiers together. Most of the existing feature transfer methods can only handle homogeneous data (i.e., data with the same dimension) and are susceptible to the quality of the RSIs, while RSIs with different resolutions present different feature dimensions and samples obtained from illumination conditions. To obtain effective classification results, unlike existing methods that focus only on the projection transformation in feature space, a joint feature-space and sample-space heterogeneous feature transfer (JFSSS-HFT) method is proposed to simultaneously process heterogeneous multi-resolution images in feature space using projection matrices of different dimensions and reduce the impact of outliers by adaptive weight factors in the sample space simultaneously to reduce the occurrence of negative transfer. Moreover, the maximum interclass variance term is embedded to improve the discriminant ability of the transferred features. To solve the optimization problem of JFSSS-HFT, the alternating-direction method of multipliers (ADMM) is introduced to alternatively optimize the parameters of JFSSS-HFT. Using different types of ship patches and airplane patches with different resolutions, the experimental results show that the proposed JFSSS-HFT obtains better classification results than the typical feature transferred methods.

Joint object recognition and pose estimation using multiple-anchor triplet learning of canonical plane

Accurate object recognition and pose estimation models are essential for practical applications of robot arms, such as picking products on a shelf. Training such a model often requires a large-scale dataset with qualified labels for both object classes and pose parameters, and collecting accurate pose labels is particularly costly. A recent paper Ueno et al.(2019)[28] proposed a triplet learning framework for joint object recognition and pose estimation without explicit pose labels by learning a spatial transformer network to estimate the pose difference of an input image from an anchor image depicting the same object in a reference pose. However, our analysis suggests that the pose estimation accuracy is severely degraded for input images with large pose differences. To address this problem, we propose a new learning approach called multiple-anchor triplet learning. The basic idea is to give dense reference poses by preparing multiple anchors so that there is at least one anchor image having a small pose difference to the input image. Our multiple-anchor triplet learning is an extension of the standard single-anchor triplet learning to the multiple-anchor case. Inspired by the idea of multiple instance learning, we introduce a selection layer that automatically chooses the best anchor for each input image and allows the network to be trained end-to-end to minimize triplet-based losses. Experiments with three benchmark datasets in product picking scenarios demonstrate that our method significantly outperforms existing methods in both object recognition and pose estimation accuracy.

Detection of building fixtures in 3D point cloud data

Building architectural and civil engineering are constantly changing, causes the increases of building spaces as well as renovation works which includes structures such as walls, ceilings and floors, and building fixtures. Building fixtures are objects which is secured to the building, such as lighting fixtures, plug and socket, ceiling fan and so on. It is considered as one of the complex structures in building as the size of the fixtures are small and sometimes are hardly seen immediately. When a certain building changes, the building information need to be updated along with the changes of the building. The process to update the changes has contributed towards complex and huge data to be processed which usually involves tedious and complicated work. Therefore, to recognize the fixtures in building environment before renovation, an object recognition method is applied. This investigation focused on the recognition of lighting fixtures in the environments. By using MATLAB, an algorithm is developed to detect the point cloud data that belongs to the lighting fixtures. The investigation shows that the lighting fixtures can be identified by using Region of Interest (ROI) method within an environment. From the results, the accuracy of the dimensions of the lighting fixtures detected in point cloud data compared to the real one in the environment is 75% and 72% match, which is good but still need an improvement to be closely match with the real dimensions. The finding is hoped to simplify the tasks of determining the fixtures in the building before any changes is done.

Decoding Brain Activity Features to Recognize Distorted Objects.

Brain decoding is able to make human interact with an external machine or robot for assisting patient's rehabilitation. Brain generic object recognition ability can be decoded through multiple neuroimaging modalities like functional magnetic resonance imaging (fMRI). On the other hand, external machine may wrongly recognize objects due to distorted noisy or blurring images caused by many factors, and therefore deteriorate performance of brain-machine interaction. In order to create better machine, generalization capability of human brain is transferred to classifier for enhancing classification accuracy of distorted images. Since homology existing between human and machine vision has been demonstrated, through decoding neural activity features of fMRI signals into feature units of convolutional neural network layers, an enhanced object recognition method is proposed to integrate brain activity into classifier for increasing classification accuracy. Experimental results show that the proposed method is able to enhance generalization capability of distorted object recognition.

Multi-view dual attention network for 3D object recognition

The existing view-based 3D object classification and recognition methods ignore the inherent hierarchical correlation and distinguishability of views, making it difficult to further improve the classification accuracy. In order to solve this problem, this paper proposes an end-to-end multi-view dual attention network framework for high-precision recognition of 3D objects. On one hand, we obtain three feature layers of query, key, and value through the convolution layer. The spatial attention matrix is generated by the key-value pairs of query and key, and each feature in the value of the original feature space branch is assigned different importance, which clearly captures the prominent detail features in the view, generates the view space shape descriptor, and focuses on the detail part of the view with the feature of category discrimination. On the other hand, a channel attention vector is obtained by compressing the channel information in different views, and the attention weight of each view feature is scaled to find the correlation between the target views and focus on the view with important features in all views. Integrating the two feature descriptors together to generate global shape descriptors of the 3D model, which has a stronger response to the distinguishing features of the object model and can be used for high-precision 3D object recognition. The proposed method achieves an overall accuracy of 96.6% and an average accuracy of 95.5% on the open-source ModelNet40 dataset, compiled by Princeton University when using Resnet50 as the basic CNN model. Compared with the existing deep learning methods, the experimental results demonstrate that the proposed method achieves state-of-the-art performance in the 3D object classification accuracy.

A Method for Automated Control of Electronic Components on Microfocus X-ray Images

Introduction. Machine vision systems are increasingly used in industrial production, particularly for monitoring the quality of electronic components. Radiographic (Х-ray) inspection is currently one of the most popular types of non-destructive testing. Electronic components are typically characterized by a small size, hence, their radiographic inspection should be based on obtaining images and their further enlargement. X-ray equipment for performing such studies is designed such that there are relatively small input doses of X-ray radiation in the plane of the receiver, which leads to a higher image noise than that using conventional X-ray devices.Aim. To develop a method for automated object recognition on microfocus X-ray images.Materials and methods. A method for segmentation of X-ray images is proposed. In the first step, adaptive median filtering is performed followed by correction of the image background by subtracting the distorting function. Next, the contours of the objects in the image are identified using the Canny edge detector followed by recognition of the objects on the resulting image.Results. The developed method was tested for quality control of the installation of microcircuits and for determining the number of electronic components. The experiments confirmed the accuracy of the proposed method. When monitoring the quality of microcircuit installation, the number of detected defects differed from that verified by the operator by less than 10 %. The average error of the proposed method was less than 0.1% when determining the number of electronic components.Conclusion. The proposed method for object recognition on microfocus X-ray images demonstrated sufficient accuracy in typical tasks of non-destructive testing of electronic components.

Superpixel Segmentation Based on Grid Point Density Peak Clustering.

Superpixel segmentation is one of the key image preprocessing steps in object recognition and detection methods. However, the over-segmentation in the smoothly connected homogenous region in an image is the key problem. That would produce redundant complex jagged textures. In this paper, the density peak clustering will be used to reduce the redundant superpixels and highlight the primary textures and contours of the salient objects. Firstly, the grid pixels are extracted as feature points, and the density of each feature point will be defined. Secondly, the cluster centers are extracted with the density peaks. Finally, all the feature points will be clustered by the density peaks. The pixel blocks, which are obtained by the above steps, are superpixels. The method is carried out in the BSDS500 dataset, and the experimental results show that the Boundary Recall (BR) and Achievement Segmentation Accuracy (ASA) are 95.0% and 96.3%, respectively. In addition, the proposed method has better performance in efficiency (30 fps). The comparison experiments show that not only do the superpixel boundaries have good adhesion to the primary textures and contours of the salient objects, but they can also effectively reduce the redundant superpixels in the homogeneous region.

A new structure of the focal element in object recognition

The research on uncertainty is a hot spot in artificial intelligence, especially in actual applications. Most research focus on measuring the uncertainty between evidence, ignoring the structure of evidence. When a sample has multiple features, the traditional structure of the focal element cannot handle and display the uncertainty between multiple features directly. Hence, we proposed a new structure of the focal element considering multiple features into one focal element and provided the corresponding basic probability assignment calculation. To apply the new structure in practices, we provided a new object recognition method combining the fractal feature and credible evidence. The new method can improve the robustness of the algorithm in a noisy environment. In the end, some experiments have illustrated the validity and correctness of the proposed method.

IDENTIFYING OIL PADS IN HIGH SPATIAL RESOLUTION AERIAL IMAGES USING FASTER R-CNN

Abstract. Deep learning (DL) methods are used for identifying objects in aerial and ground-based images. Detecting vehicles, roads, buildings, and crops are examples of object identification applications using DL methods. Identifying complex natural and man-made features continues to be a challenge. Oil pads are an example of complex built features due to their shape, size, and presence of other structures like sheds. This work applies Faster Region-based Convolutional Neural Network (R-CNN), a DL-based object recognition method, for identifying oil pads in high spatial resolution (1m), true-color aerial images. Faster R-CNN is a region-based object identification method, consisting of Regional Proposal Network (RPN) that helps to find the area where the target can be possibly present in the images. If the target is present in the images, the Faster R-CNN algorithm will identify the area in an image as foreground and the rest as background. The algorithm was trained with oil pad locations that were manually annotated from orthorectified imagery acquired in 2017. Eighty percent of the annotated images were used for training and the number of epochs was increased from 100 to 1000 in increments of 100 with a fixed length of 1000. After determining the optimal number of epochs the performance of the algorithm was evaluated with an independent set of validation images consisting of frames with and without oil pads. Results indicate that the Faster R-CNN algorithm can be used for identifying oil pads in aerial images.

Review of multi-view 3D object recognition methods based on deep learning

Three-dimensional (3D) object recognition is widely used in automated driving, medical image analysis, virtual/augmented reality, artificial intelligence robots, and other areas. Deep learning is increasingly being used to solve 3D vision problems. Multi-view 3D object recognition based on the deep learning technique has become one of the rigorously researched topics because it can directly use the pretrained and successful advanced classification network as the backbone network, and views from multiple viewpoints can complement each other’s detailed features of the object. However, some challenges still exist in this area. Recently, many methods have been proposed to solve the problems pertaining to this research topic. This paper presents a comprehensive review and classification of the latest developments in the deep learning methods for multi-view 3D object recognition. It also summarizes the results of these methods on a few mainstream datasets, provides an insightful summary, and puts forward enlightening future research directions.

Object Recognition and Grasping Detection Method of Construction Machinery Robot Relying on Deep Learning

Object Recognition and Grasping Detection Method of Construction Machinery Robot Relying on Deep Learning

Fault Maintenance and Repair Application Based on Augmented Reality Technology in Electrical Distribution Systems

This study presents an AR technology-based software that will work on mobile platforms and the pilot application of this software to solve the possible problems that may be experienced other than routine maintenance activities on protection/control panels and devices in SCADA centers. AR is essentially an information technology that provides the ability to present digital data and images in the same environment as our real world through its visualization, instruction/guidance, and interaction capabilities. Object recognition methods, necessary software, and infrastructure requirements are mentioned in this study in order to benefit from these advantages of AR technology correctly. Different scenarios have been created by using AR for possible failures in automation panels in a distribution center with a telemetry system. Finally, an exemplary scenario in the study is realized in a distribution center located in the SEDAŞ area of responsibility.

POS0036 AN ARTIFICIAL INTELLIGENCE MODEL IN RHEUMATOLOGY: INTERPRETATION OF THE SACROILIAC JOINT GRAPHY IN ANKYLOSING SPONDYLITIS

Background:There is an average of 8 years delay in the diagnosis of ankylosing spondylitis (AS). The most important danger of late diagnosis is that the disease can cause physical and functional disability (2). There is no specific diagnostic biomarker for AS. Sacroiliac joint (SIJ) radiography is frequently used in the diagnosis and follow-up of AS due to its easy accessibility and low cost. It can be classified as grade 0, 1, 2, 3, 4, and these classes may not be sharply separated from each other (3).Objectives:Interpretation of the SIJ radiography may differ from physician to physician. In fact, the same physician may interpret it differently at different times (3). We wanted to find a solution to the intraobserver disagreement problem with the artificial intelligence model.Methods:The SIJ radiography of 590 patients who applied to our center were divided into 3 categories as right and left, separately, grade 0, grade 1-2, grade 3-4, and an educational data set was prepared for the object recognition method. 488 images were augmented through noise from 490 images in the training data. 242 articular objects were trained for grade 0, 278 for grade 1-2, and 1426 for grade 2-3. The model was tested with 100 images for 36 joint objects for grade 0, 29 for grade 1-2, and 135 for grade 3-4 to create a computer vision-artificial intelligence model (image 1).Results:Training performance is 70% for grade 0, %63 for grade 1-2, %90 for grade 3-4 and test performance is %52 for grade 0, %24 for grade 1-2, %86 intersection over union (I/U:Intersection over Union is a form of measurement used to indicate the accuracy of an object detector.) for grade 3-4. The mean average precision (mAP) score of our object detection model is %65.9 for test data set (image 1). The estimation quality of the model can be affected by the distribution and number of each class.Conclusion:The experience of the x-ray technician, dose adjustment, and position differences due to patient compliance complicate the standardization of SIJ radiography and this may cause interobserver disagreement (3). Artificial intelligence models to be created with a larger and homogeneous data set in order to ensure objective standardization in the interpretation of the SIJ graph can help physicians.

Video Surveillance System of Target Contour

The paper considers the problem of object detection and recognition, namely, its solution without the use of expensive, resource-intensive and complex data collection and processing systems, with the possibility of its mobility, ease of installation and initial setup. The available ways and methods of solving the problem, their advantages and disadvantages are given. The operation of the system according to an algorithm developed on the basis of the method of object recognition, namely the selection of contours by a filter based on the Canny operator, is presented. The article presents intermediate and final illustrations of the system algorithm at each step, which gives an opportunity to get acquainted with its advantages over classical video surveillance systems and some disadvantages. Elements such as a webcam with a video frame rate of 25 frames per second, a mobile phone and a personal computer with the MatLab2021 programming environment installed (chosen due to the ease of use and the presence of built-in image processing functions) were used to demonstrate the system's performance.

ПРИМЕНЕНИЕ НЕЙРОННЫХ СЕТЕЙ ДЛЯ РАСПОЗНАВАНИЯ ПРИНЦИПИАЛЬНЫХ УСЛОВНО-ГРАФИЧЕСКИХ ЭЛЕКТРИЧЕСКИХ ОБОЗНАЧЕНИЙ

The article discusses the issue of using artificial neural networks for recognizing the conditionally graphical designations of electrical engineering, in particular, the convolutional neural networks and the R-CNN object recognition model, which is most suitable for solving the task at hand. Recognition of images of a specific picture is a task set for the complex information processing systems, as well as control and decision-making systems. The classification of various technological or natural objects, analog and digital signals is developed by a set of specific characteristics and properties. Defining the type and features of an object finds its application in different branches of science: machine learning, diagnostics, meteorology, video surveillance and security systems, in virtual reality systems and image search. However, research has not yet been carried out for solving the applied problems and achieving the required parameters (e.g. in recognizing conditional graphical symbols of electrical engineering). The neural networks have been found to have the highest quality and most promising among all mathematical models and methods of pattern recognition. As for the interactivity, the output result of image recognition work is a necessary and sufficient answer, which does not have a stable work on the variability of objects within categories and their invariant transformations. The scheme of the model R-CNN has been studied in detail, the importance of the training sample and its influence on the quality of pattern recognition by the neural network have been grounded. The application of the RoI Pooling method for object recognition in the image is shown in general, due to which there have been selected several regions of interest indicated through the bounding boxes.

Artificial and beneficial – Exploiting artificial images for aerial vehicle detection

Object detection in aerial images is an important task in environmental, economic, and infrastructure-related tasks. One of the most prominent applications is the detection of vehicles, for which deep learning approaches are increasingly used. A major challenge in such approaches is the limited amount of data that arises, for example, when more specialized and rarer vehicles such as agricultural machinery or construction vehicles are to be detected. This lack of data contrasts with the enormous data hunger of deep learning methods in general and object recognition in particular. In this article, we address this issue in the context of the detection of road vehicles in aerial images. To overcome the lack of annotated data, we propose a generative approach that generates top-down images by overlaying artificial vehicles created from 2D CAD drawings on artificial or real backgrounds. Our experiments with a modified RetinaNet object detection network show that adding these images to small real-world datasets significantly improves detection performance. In cases of very limited or even no real-world images, we observe an improvement in average precision of up to 0.70 points. We address the remaining performance gap to real-world datasets by analyzing the effect of the image composition of background and objects and give insights into the importance of background.

Local feature descriptor using discrete first and second fundamental forms

3D object recognition from point clouds is a fast-growing field of research in computer vision. 3D object recognition methods can be classified into two categories: global feature-based and local feature-based methods. The local feature-based methods are more popular than global ones. Because the global feature-based methods need a prior segmentation of the scene, they are not suitable for real-world scenes. Many previous local descriptor methods limit their performance by introducing a local reference frame or axis (LRF/A). Estimating the LRF/A for each keypoint leads to extra computational time and error. We use the fundamental theorem of surface theory to introduce a simple and efficient local feature descriptor based on the coefficients of discrete first and second fundamental forms. The proposed method overrides the necessity of an LRF/A, and it required a small feature dimension of seven, which means it is a low-complexity and fast procedure. To assess the proposed method, we have compared it with eight state-of-the-art descriptors and applied it to the three popular datasets to extract features and recognize the correspondences. Experimental results demonstrate the superiority of the proposed approach to the compared methods in terms of pairwise registration measurements, recall versus 1-precision curve, and the computational time.