Object Recognition Scheme Research Articles

Optimization Method for Robot Moving Object Recognition and Grasping Strategy Based on Binocular Vision

<p>This article proposes a more accurate grasping strategy for the recognition and grasping of moving targets based on binocular vision cameras. Firstly, the front and back scene separation algorithm is used to identify the moving target grabbing object in the production line. Then, by setting an appropriate threshold, the SiamMask target tracking algorithm is improved to achieve dynamic target tracking. Finally, the conveyor belt speed is detected and the real-time position of the object is obtained. Then, the Cartesian strategy is used to achieve path planning and optimization methods for the robotic arm during movement. Through experimental simulation, the effectiveness and stability of the proposed method in this paper have been demonstrated.</p> <p> </p>

Object Recognition Scheme for Digital Transformation in Marine Science and Engineering

With the advancement of deep learning (DL), researchers and engineers in the marine industry are exploring the application of DL technologies to their specific applications. In general, the accuracy of inference using DL technologies is significantly dependent on the number of training datasets. Unfortunately, people in marine science and engineering environments are often reluctant to share their documents (i.e., P&ID) with third-party manufacturers or public clouds to protect their proprietary information. Despite this, the demand for object detection using DL technologies in image-formatted files (i.e., jpg, png, or pdf format) is steadily growing. In this paper, we propose a new mechanism, called a no-training object picker (NoOP), which efficiently recognizes all objects (e.g., lines, tags, and symbols) in image-formatted P&ID documents. Notably, it can recognize objects without any training dataset, thus reducing the time and effort required for training and collection of unpublished datasets. To clearly present the effectiveness of NoOP, we evaluated NoOP using a real P&ID document. As a result, we confirmed that all objects in the image-formatted P&ID file are successfully detected over a short time (only 7.11 s on average).

Predictive tracking of an object by a pan–tilt camera of a robot

Moving-object tracking using a pan–tilt camera setup is quite a well-known task in robotics. However, the presented research addresses specific properties of the tracked object and introduces novel features to the pan–tilt camera control strategy. Pan–tilt camera control does not operate in an isolated environment. It is a part of the visual servoing system with specific goals. The system has to fulfill certain purposes, which affect its configuration and functionality. The pan–tilt system aims at keeping the visually tracked object within the middle of the image. At the same time, the overall visual servoing efficiently recognizes and tracks the object enabling its grasping by the robot arm. It uses a predictive strategy utilizing specific second-order linear models for pan and tilt joints. Model predictive control (MPC) introduces into the system the ability to predict camera operation over the specific horizon according to the predefined tracking goals. As the system anticipates future positions over the horizon of operation, the setpoint prediction of the future tracked system positions is required. Visual object recognition and tracking system use particular strategies for preparing online tracked object extrapolation over MPC horizon. Therefore, the pan–tilt camera system is intrinsically coupled to camera-based recognition and tracking. Predictive pan–tilt positioning keeps the tracked system in the middle of the image, while the visual system extrapolation improves the tracking performance. The proposed approach is thoroughly tested in the dedicated Gazebo-based robot simulator. Finally, the system is implemented and validated on the Velma robot. The results and their comparison with other control strategies confirm the initial assumptions, allowing further visual servoing system development.

Rats use strategies to make object choices in spontaneous object recognition tasks

Rodent spontaneous object recognition (SOR) paradigms are widely used to study the mechanisms of complex memory in many laboratories. Due to the absence of explicit reinforcement in these tasks, there is an underlying assumption that object exploratory behaviour is ‘spontaneous’. However, rodents can strategise, readily adapting their behaviour depending on the current information available and prior predications formed from learning and memory. Here, using the object-place-context (episodic-like) recognition task and novel analytic methods relying on multiple trials within a single session, we demonstrate that rats use a context-based or recency-based object recognition strategy for the same types of trials, depending on task conditions. Exposure to occasional ambiguous conditions changed animals’ responses towards a recency-based preference. However, more salient and predictable conditions led to animals exploring objects on the basis of episodic novelty reliant on contextual information. The results have important implications for future research using SOR tasks, especially in the way experimenters design, analyse and interpret object recognition experiments in non-human animals.

Multi-modal cyber security based object detection by classification using deep learning and background suppression techniques

Foreground object recognition in video is critical in many computer vision applications and automated video surveillance systems. Object detection and tracking are critical steps in navigation, object recognition and surveillance schemes. Object detection is process of separating foreground and background items in photographs. In this paper, we proposes a framework for achieving these tasks with the enhanced cyber security control facility. This proposed algorithmperformed in 2 stages: multi-object detection utilizing the Cyber secure Probabilistic Gaussian Mixture Model (Cy_SPGMM) and background suppression and another stage is multiple moving objects tracking utilizing Kernel convoluted moving window with Kalman filter (KCMW_KF). It can, however, deal with a variety of video sequences in the MOT 20 dataset. The experimental findings reveal that the proposed algorithm detects and tracks foreground objects in complex and dynamic scenarios with high accuracy, robustness, and efficiency. This method also produces smoothened images without noise.

HS‐IDEOR—A tool for conceptual modeling and designing of hardware architectures focused on object recognition

AbstractConsidering the increasing demand for object recognition systems, it is crucial that modern engineering courses include tools and methods to ensure that students acquire the theoretical and practical skills necessary to become specialists in designing such systems. This paper introduces a hardware‐software integrated development environment to support and accelerate the teaching/learning process of conceptual modeling and hardware architectures design focused on object recognition. In addition, our proposal facilitates the implementation of modular and hybrid object recognition schemes, offering great flexibility in the study of such systems. An empirical evaluation was conducted with 30 fourth‐year undergraduate students who designed and implemented an object recognition system in the Computer Vision course of the Bachelor's Degree in Computer Science and Engineering at the Universidad Tecnológica de la Mixteca, Mexico. The results achieved from designing, modeling, and implementing a hardware architecture of the convolution process applied to edge detection in an image provided evidence that this approach empowered students with a more dynamic learning process.

Neurocomputing for internet of things: Object recognition and detection strategy

Modern and new integrated technologies have changed the traditional systems by using more advanced machine learning, artificial intelligence methods, new generation standards, and smart and intelligent devices. The new integrated networks like the Internet of Things (IoT) and 5G standards offer various benefits and services. However, these networks have suffered from multiple object detection, localization, and classification issues. Conventional Neural Networks (CNN) and their variants have been adopted for object detection, classification, and localization in IoT networks to create autonomous devices to make decisions and perform tasks without human intervention and helpful to learn in-depth features. Motivated by these facts, this paper investigates existing object detection and recognition techniques by using CNN models used in IoT networks. This paper presents a Conventional Neural Networks for 5G-Enabled Internet of Things Network (CNN-5GIoT) model for moving and static objects in IoT networks after a detailed comparison. The proposed model is evaluated with existing models to check the accuracy of real-time tracking. The proposed model is more efficient for real-time object detection and recognition than conventional methods.

Fs-DSM: Few-Shot Diagram-Sentence Matching via Cross-Modal Attention Graph Model.

Diagram-sentence matching is a valuable academic research because it can help learners effectively understand the diagrams with the assisted by sentences. However, there are many uncommon objects, i.e. few-shot contents in diagrams and sentences. The existing methods for image-sentence matching have great limitations when applied to diagrams. Because they focus on the high-frequency objects during training and ignore the uncommon objects. In addition, the specialty leads to the semantic non-intuition of the diagram itself. In this work, we propose a cross-modal attention graph model for the few-shot diagram-sentence matching task named Fs-DSM. Specifically, it is composed of three modules. The graph initialization module regards the region-level diagram features and word-level sentence features as the nodes of Fs-DSM, and edges are represented as similarity between nodes. The information propagation module is a key point of Fs-DSM, in which the few-shot contents are recognized by an uncommon object recognition strategy, and then the nodes are updated by a neighborhood aggregation procedure with cross-modal propagation between all visual and textual nodes, while the edges are recomputed based on the new node features. The global association module integrates the features of regions and words to represent the global diagrams and sentences. By conducting comprehensive experiments in terms of few-shot and conventional image-sentence matching, we demonstrate that Fs-DSM achieves superior performances over the competitors on the AI2D [Formula: see text] diagram dataset and two public benchmark datasets with nature images.

Tensor-Train decomposition for image recognition

We explore the potential of Tensor-Train (TT) decompositions in the context of multi-feature face or object recognition strategies. We devise a new recognition algorithm that can handle three or more way tensors in the TT format, and propose a truncation strategy to limit memory usage. Numerical comparisons with other related methods—including the well established recognition algorithm based on high-order SVD—illustrate the features of the various strategies on benchmark datasets.

What Versus How in Visual Search: Effects of Object Recognition Training, Strategy Training, and Non-invasive Brain Stimulation on Satellite Image Search

Visual search is required in many professions where an undetected threat, such as a weapon, can put the well-being of others at risk. Given the importance of detecting these threats, researchers have used various experimental techniques to improve performance in visual search tasks, albeit with varying degrees of success. Here, we explore two promising techniques to improve visual search using ecologically valid synthetic aperture radar stimuli: object recognition training and search strategy training. Search strategy training is intended to make observers search more systematically through a display, whereas object recognition training is intended to improve observers’ ability to recognize critical targets. Search strategy training was implemented by instructing participants to scan through the display in a pre-specified pattern. Object recognition training was implemented by having participants discriminate between targets and non-targets. We also manipulated whether observers received anodal or sham transcranial direct current stimulation (tDCS) during training, which has been shown to improve visual search performance and target learning. To measure the effectiveness of the training and stimulation conditions, we tested object recognition accuracy and overall visual search performance before and after three sessions of increasingly difficult training. Results indicated that object recognition training significantly improved object recognition accuracy relative to the search strategy group, whereas search strategy training was effective in improving visual search accuracy in those who adhered to the training. However, tDCS did not interact with training type, and although both training types yielded significant improvements, training-related improvements were not significantly different between the different approaches. This evidence suggests that strategy-based training could be as effective as the more prototypical object recognition training. Implications for future training protocols are discussed.

VANET based self-governed vehicle strategy in association with multi-object recognition and key mapping features

Vehicular technologies are the most important and well-known communication strategy, which enables vehicle to vehicle communication, inter and intra vehicle communication abilities to users to achieve their communication needs. Generally vehicular technologies are acquired from the logics of mobile communication technologies, apart from that the vehicular logics are quite different in the sense of road safety measures but the mobile communication mode concentrates only on communication strategies. Last few decades these vehicular communication medium analysts give more concern on Self-Governed vehicles, which are frequently known as autonomous-vehicles. This intelligent invention assist vehicles with own brain and run the vehicles on road without any human interventions and also this kind of feature raises many questions to analysts such as: how safety norms are concerned over this approach, what are all the safety precautions are required to make this prototypical model to real-time application and so on. These different kind of queries raise huge pressure to analysts to derive this application. In this paper, a new intellectual approach is designed to provide a solution to vehicular technology issues specifically self-governed vehicular logics, which is called as Self-Governed Intellectual Vehicular Model (SGIVM). The proposed algorithm SGIVM provides a solution to road safety measures along with intelligent traffic manipulations by means of two different methodologies such as Object Recognition Strategy and Key Mapping on road sides/directions. With these two features, a new Self-Governed algorithm is designed and this paper assures best accuracy in results with following parameters such as safety measures, object recognition accuracy, mapping accuracy and vehicle speed maintenance level. The proposed result will prove all these features such as self-governed vehicular medium in association with object recognition and key mapping facility enables good results in outcome and it will be described / proved in detail with following summaries.

Onboard Real-time Object Surface Recognition for a Small Indoor Mobile Platform Based on Surface Component Ratio Histogram

Since a RGB-D sensor provides rich information about the scene, various object recognition schemes and low-level image descriptors can be used to improve the SLAM performance. However, a cleaning robot, which is usually flat and thus the camera is close to the floor, usually only has a partial view of the objects in front of the camera; therefore, conventional object recognition schemes based on the complete view of objects are not suitable. To address this problem, we introduce a novel object surface recognition algorithm based on the proposed surface component ratio histogram (SCRH). SCRH is a surface descriptor which describes the geometrical shape of the partial view of the object. Without any pre-trained model of the objects, SCRH provides a way to deal with the unexpected objects which the robot encounters during the navigation. SCRH was evaluated using several objects lying on the floor of which the identities are not known in advance. The experimental results show that objects are successfully discriminated based on their surfaces and SCRH is robust for object surface recognition.

Enet-CRF-Lidar: Lidar and Camera Fusion for Multi-Scale Object Recognition

Object recognition, which mainly includes object detection and semantic segmentation, is one of the critical challenges for intelligent vehicles. In most cases, cameras and Lidar are the most common sensors used for object recognition. However, both cameras and Lidar suffer from some inherent drawbacks. Therefore, the fusion of camera and Lidar becomes a natural solution to overcome the inherent defects of each single sensing modality. With the boost of deep learning-based algorithms, multi-sensor fusion methodologies employ deep-learning methods as their fusion strategy, which has made impressive accomplishments on large-scale objects such as vehicles and buses. However, most existing sensor-fusion strategies have the problem of ignoring detailed information caused by down-sampling operations in deep learning, which results in poor detection performance on small-scale objects such as pedestrians and cyclists. In this paper, we propose a real-time multi-sensor (Lidar and color camera) fusion strategy for multi-scale object recognition at the semantic level named Enet-CRF-Lidar. Firstly, a multi-module Enet is designed to adapt both large-scale objects and small-scale ones. Then, the CRF-RNN module is integrated with the multi-module Enet to introduce the low-level details of the input data, which leads to a significant improvement in small-scale object recognition. The experimental results show that the proposed Enet-CRF-Lidar module can provide reliable detection performance on multi-scale objects and can be adapted to complex scenarios.

A 20-ch TDC/ADC Hybrid Architecture LiDAR SoC for 240 <inline-formula> <tex-math notation="LaTeX">$\times$ </tex-math> </inline-formula> 96 Pixel 200-m Range Imaging With Smart Accumulation Technique and Residue Quantizing SAR ADC

This paper presents a time-to-digital converter/analog-to-digital-converter (TDC/ADC) hybrid LiDAR system-on-chip (SoC) to realize reliable self-driving systems. The smart accumulation technique (SAT) is proposed to achieve both 200-m and high-pixel-resolution range imaging, which was untrodden with conventional LiDARs. The “smart” accumulation is realized by a simple object recognition strategy with small circuit overhead. When compared to conventional accumulations, the LiDAR range is enhanced without degrading the pixel resolution. Moreover, a TDC/ADC hybrid architecture is proposed to achieve a wide-distance-range LiDAR with a small silicon area and short-range precision. To minimize the ADC cost, a residue-quantizing noise-shaping (RQNS) SAR ADC is proposed. The prototype LiDAR SoC is fabricated in the 28-nm CMOS technology and integrated into the silicon photomultiplier (SiPM)-based LiDAR system. LiDAR measured with 240 $\times$ 96 pixels at 10 frames/s achieves a measurement range of 200 m with a 70-klx direct sunlight: the measurement range is 2 $\times$ longer than conventional designs. Furthermore, our LiDAR achieves 4 $\times$ higher effective pixel resolution compared to conventional designs using simple accumulation. A 3-D point-cloud image acquired with a real-life environment is presented.

Tactile-based active object discrimination and target object search in an unknown workspace

The tasks of exploring unknown workspaces and recognizing objects based on their physical properties are challenging for autonomous robots. In this paper, we present strategies solely based on tactile information to enable robots to accomplish such tasks. (1) An active exploration approach for the robot to explore unknown workspaces; (2) an active touch objects learning method that enables the robot to learn efficiently about unknown objects via their physical properties (stiffness, surface texture, and center of mass); and (3) an active object recognition strategy, based on the knowledge the robot has acquired. Furthermore, we propose a tactile-based approach for estimating the center of mass of rigid objects. Following the active touch for workspace exploration, the robotic system with the sense of touch in fingertips reduces the uncertainty of the workspace up to 65 and 70% compared respectively to uniform and random strategies, for a fixed number of samples. By means of the active touch learning method, the robot achieved 20 and 15% higher learning accuracy for the same number of training samples compared to uniform strategy and random strategy, respectively. Taking advantage of the prior knowledge obtained during the active touch learning, the robot took up to 15% fewer decision steps compared to the random method to achieve the same discrimination accuracy in active object discrimination task.

Vanets Meet Autonomous Vehicles: Multimodal Surrounding Recognition Using Manifold Alignment

In the past two years, calls for developing synergistic links between the two worlds of vehicular ad-hoc networks (VANETs) and autonomous vehicles have significantly gone up to achieve further on-road safety and benefits for end-users. In this paper, we present our vision to create such a beneficial link by designing a multimodal scheme for object detection, recognition, and mapping based on the fusion of stereo camera frames, point cloud Velodyne LIDAR scans, and vehicle-to-vehicle (V2V) basic safety messages (BSMs) exchanges using VANET protocols. Exploiting the high similarities in the underlying manifold properties of the three data sets, and their high neighborhood correlation, the proposed scheme employs semi-supervised manifold alignment to merge the key features of rich texture descriptions of objects from 2-D images, depth and distance between objects provided by 3-D point cloud, and the awareness of self-declared vehicles from BSMs’ 3-D information including the ones not seen by camera and LIDAR. The proposed scheme is applied to create joint pixel-to-point-cloud and pixel-to-V2V correspondences of objects in frames from the KITTI Vision Benchmark Suite, using a semi-supervised manifold alignment, to achieve camera-LIDAR and camera-V2V mapping of their recognized objects. We present the alignment accuracy results over two different driving sequences and show the additional acquired knowledge of objects from the various input modalities. We also study the effect of the number of neighbors employed in the alignment process on the alignment accuracy. With proper choice of parameters, the testing of our proposed scheme over two entire driving sequences exhibits 100% accuracy in the majority of cases, 74%–92% and 50%–72% average alignment accuracy for vehicles and pedestrians and up to 150% additional object recognition of the testing vehicle’s surrounding.

Object recognition with discrete orthogonal Hahn moments

In this research, we have analysed the image feature descriptive properties of discrete orthogonal Hahn moments, and proposed a new object recognition scheme with three modes of Hahn moment descriptors, global, local, and hybrid, respectively. For each mode, we have employed the four highest variances values from ten lowest order of Hahn moments to compose a four-dimensional feature vector. To clarify our new scheme of using discrete orthogonal Hahn moment characteristics, we utilised a set of 6,763 Chinese characters defined in China's national standard GB2312, with the font of song, as the testing object set. Each of the three Hahn moment modes has performed very well, while the experimental results of utilising the three Hahn moment modes are independent from each other.

Object recognition with discrete orthogonal Hahn moments

In this research, we have analysed the image feature descriptive properties of discrete orthogonal Hahn moments, and proposed a new object recognition scheme with three modes of Hahn moment descriptors, global, local, and hybrid, respectively. For each mode, we have employed the four highest variances values from ten lowest order of Hahn moments to compose a four-dimensional feature vector. To clarify our new scheme of using discrete orthogonal Hahn moment characteristics, we utilised a set of 6,763 Chinese characters defined in China's national standard GB2312, with the font of song, as the testing object set. Each of the three Hahn moment modes has performed very well, while the experimental results of utilising the three Hahn moment modes are independent from each other.

Improved Object Recognition with Decision Trees Using Subspace Clustering

Generic object recognition algorithms usually require complex classificationmodels because of intrinsic difficulties arising from problems such as changes in pose, lighting conditions, or partial occlusions. Decision trees present an inexpensive alternative for classification tasks and offer the advantage of being simple to understand. On the other hand, a common scheme for object recognition is given by the appearances of visual words, also known as the bag-of-words method. Although multiple co-occurrences of visual words are more informative regarding visual classes, a comprehensive evaluation of such combinations is unfeasible because it would result in a combinatorial explosion. In this paper, we propose to obtain the multiple co-occurrences of visual words using a variant of the CLIQUE subspace-clustering algorithm for improving the object recognition performance of simple decision trees. Experiments on standard object datasets show that our method improves the accuracy of the classification of generic objects in comparison to traditional decision tree techniques that are similar, in terms of accuracy, to ensemble techniques. In future we plan to evaluate other variants of decision trees, and apply other subspace-clustering algorithms.

Automatic anatomy recognition in whole-body PET/CT images.

Whole-body positron emission tomography/computed tomography (PET/CT) has become a standard method of imaging patients with various disease conditions, especially cancer. Body-wide accurate quantification of disease burden in PET/CT images is important for characterizing lesions, staging disease, prognosticating patient outcome, planning treatment, and evaluating disease response to therapeutic interventions. However, body-wide anatomy recognition in PET/CT is a critical first step for accurately and automatically quantifying disease body-wide, body-region-wise, and organwise. This latter process, however, has remained a challenge due to the lower quality of the anatomic information portrayed in the CT component of this imaging modality and the paucity of anatomic details in the PET component. In this paper, the authors demonstrate the adaptation of a recently developed automatic anatomy recognition (AAR) methodology [Udupa et al., "Body-wide hierarchical fuzzy modeling, recognition, and delineation of anatomy in medical images," Med. Image Anal. 18, 752-771 (2014)] to PET/CT images. Their goal was to test what level of object localization accuracy can be achieved on PET/CT compared to that achieved on diagnostic CT images. The authors advance the AAR approach in this work in three fronts: (i) from body-region-wise treatment in the work of Udupa et al. to whole body; (ii) from the use of image intensity in optimal object recognition in the work of Udupa et al. to intensity plus object-specific texture properties, and (iii) from the intramodality model-building-recognition strategy to the intermodality approach. The whole-body approach allows consideration of relationships among objects in different body regions, which was previously not possible. Consideration of object texture allows generalizing the previous optimal threshold-based fuzzy model recognition method from intensity images to any derived fuzzy membership image, and in the process, to bring performance to the level achieved on diagnostic CT and MR images in body-region-wise approaches. The intermodality approach fosters the use of already existing fuzzy models, previously created from diagnostic CT images, on PET/CT and other derived images, thus truly separating the modality-independent object assembly anatomy from modality-specific tissue property portrayal in the image. Key ways of combining the above three basic ideas lead them to 15 different strategies for recognizing objects in PET/CT images. Utilizing 50 diagnostic CT image data sets from the thoracic and abdominal body regions and 16 whole-body PET/CT image data sets, the authors compare the recognition performance among these 15 strategies on 18 objects from the thorax, abdomen, and pelvis in object localization error and size estimation error. Particularly on texture membership images, object localization is within three voxels on whole-body low-dose CT images and 2 voxels on body-region-wise low-dose images of known true locations. Surprisingly, even on direct body-region-wise PET images, localization error within 3 voxels seems possible. The previous body-region-wise approach can be extended to whole-body torso with similar object localization performance. Combined use of image texture and intensity property yields the best object localization accuracy. In both body-region-wise and whole-body approaches, recognition performance on low-dose CT images reaches levels previously achieved on diagnostic CT images. The best object recognition strategy varies among objects; the proposed framework however allows employing a strategy that is optimal for each object.