Multi Objects Research Articles

GHE-Ensemble: Enhanced Hybrid Image Enhancement Model for Night Vision Multi-Object detection in Autonomous Vehicle

Autonomous vehicle is the exponential growing topic in the research filed due its huge demand in the current industry. Where object detection plays the major role to take any major decision. Object detection in the normal vision yields the excellent performance due to deep learning models. But object detection in the night vision does not yield the good performance due to various challenges in it. In the proposed work created a night vision database called BDD-Darko where it contains night vision images with multi objects. Applied novel GHE-Ensemble model for image enhancement technique and trained using deep learning model Yolov5 for the multi-object detection in the night vision. Which resulted in 65.3 % accuracy. Proposed model is yielding better results for image enhancement technique for night vision than the previous existing model and detecting the multi object in the single frame with better performance. This paper proposes (1)A new database called as BDD-Darko which contains night vision image with multiple objects. (2) A Novel GHE-Ensemble model for image enhancement technique for night vision images

Development of a 3-DOF rotary piezoelectric actuator with compact structure

Piezoelectric actuators have been widely applied in precision electromechanical equipment due to their high resolution, fast response, and flexible structure. However, existing rotary piezoelectric actuators still face the contradiction between the structure miniaturization and the multi degrees of freedom (MDOF) motion output. Therefore, a novel rotary piezoelectric actuator is developed to achieve characteristics of compact structure and 3-DOF motions in this study, in which an orthogonal axis support system for a spherical rotor is adopted. And a stick-slip combined direct excitation scheme is designed to obtain large working range and microradian-resolution outputs. The configuration and motion principle are designed and illustrated in detail. The statics model is established for structural optimization design and parameters determination. The actuator prototype is fabricated and tested, a compact structure is achieved with an overall size of 72 mm in diameter and 80 mm in height. 3-DOF full circle rotary motion outputs around axes X, Y and Z are achieved, the corresponding maximum rotary speeds are 0.423, 0.438, and 0.975 rad/s, and the corresponding resolutions are 4.14, 4.28 and 2.32 μrad, respectively. Moreover, the prototype has successfully completed the imaging application of multi objects in large working range. The experimental results preliminarily demonstrate the potential application of the 3-DOF rotary piezoelectric actuator in the field requiring large working range and microradian-resolution.

Adaptive Weighted Flow Net Algorithm for Human Activity Recognition Using Depth Learned Features

Human Activity Recognition (HAR) from video data collections is the core application in vision tasks and has a variety of utilizations including object detection applications, video-based behavior monitoring, video classification, and indexing, patient monitoring, robotics, and behavior analysis. Although many techniques are available for HAR in video analysis tasks, most of them are not focusing on behavioral analysis. Hence, a new HAR system analysis the behavioral activity of a person based on the deep learning approach proposed in this work. The most essential aim of this work is to recognize the complex activities that are useful in many tasks that are based on object detection, modelling of individual frame characteristics, and communication among them. Moreover, this work focuses on finding out the human actions from various video resolutions, invariant human poses, and nearness of multi objects. First, we identify the key and essential frames of each activity using histogram differences. Secondly, Discrete Wavelet Transform (DWT) is used in this system to extract coefficients from the sequence of key-frames where the activity is localized in space. Finally, an Adaptive Weighted Flow Net (AWFN) algorithm is proposed in this work for effective video activity recognition. Moreover, the proposed algorithm has been evaluated by comparing it with the existing Visual Geometry Group (VGG-16) convolution neural networks for making performance comparisons. This work focuses on competent deep learning-based feature extraction to discriminate the activities for performing the classification accuracy. The proposed model has been evaluated with VGG-16 using a combination of regular UCF-101 activity datasets and also in very challenging Low-quality videos such as HMDB51. From these investigations, it is proved that the proposed AWFN approach gives higher detection accuracy of 96%. It is approximately 0.3% to 7.88% of higher accuracy than state-of-art methods.

Tactile Sensor Array With Tertiary Leaf-Vein Structures and Position-Encoded Capacity

Flexible tactile sensor is a preferred alternative in the development of human-machine interface (HMI) and robotics, which directly influences how we understand and interact with physical world. High density and high sensitivity tactile sensor array are urgently desired to assist people acquire tactile information of multi objects. However, although the existing researches exhibit great perception of tactile signals with different magnitudes and frequencies, they also expose limitation in the retrieval of relative position relationships. Herein, we report a wearable and high density (96 sensors on the fingertip) tactile sensor array based on tertiary leaf-vein inspired piezoresistive film. Particularly, a tertiary microstructure in the direction of height, including mid-vein-mimicking wrinkle (MVW), lateral-vein-mimicking wrinkles (LVW) and minor-vein-mimicking wrinkles (mVW), has been demonstrated an effective method to fabricate rough surface. The highly surface-controllable piezoresistive film dramatically enhances the durability (over 5000 cycles) and improves the sensitivity from 8.44 kPa <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">-1</sup> to 25.80 kPa <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">-1</sup> under 2 kPa. Accordingly, we propose an innovative position-encoded algorithm aimed at high density tactile sensor array, which has been demonstrated an excellent encode capacity of translating the multi-location tactile information caused by Braille characters into human-readable outputs. This study not only provides a map scanning method for Braille recognition, but also has significant implications for position-based information exchange or tactile encryption.

Ontology based multiobject segmentation and classification in sports videos

The primary objective is to identify and segments the multiple, partly occluded objects in the image. The subsequent stage carry out our approach, primarily start with frame conversion. Next in the preprocessing stage, the Gaussian filter is employed for image smoothening. Then from the preprocessed image, Multi objects are segmented through modified ontology-based segmentation, and the edge is detected from the segmented images. After that, from the edge detected frames area is extracted, which results in object detected frames. In the feature extraction stage, attributes such as area, contrast, correlation, energy, homogeneity, color, perimeter, circularity are extorted from the detected objects. The objects are categorized as human or other objects (bat/ball) through the feed-forward back propagation neural network classifier (FFBNN) based upon the extracted attributes.

Natural scene text detection based on YOLO V2 network model

As an important part in the field, the natural scene text detection has been widely applied in visual navigation system, content-based image and video retrieval, instant translation system and so on. In this paper, we introduce several object detection network based on deep learning, and apply the YOLO v2 into natural scene text detection, changing the multi objects detection problems into the two classification problems. The main works in the paper include the following: prepare the datasets; we train the YOLO v2 with the optimum parameters, carry out the regression analysis of the coordinate parameters and categories of bounding boxes, obtain the detection result; according to different detection models, the detection results of different datasets are compared and analyzed, YOLO V2 model detection speed 0.105s/image has certain advantages.

TECHNIQUE OF C PROGRAM TRANSLATION FOR RECONFIGURABLE AND HYBRID COMPUTER SYSTEMS BASED ON FIELD-PROGRAMMABLE GATE ARRAYS

In this paper, we thoroughly consider the technique of conversion of procedural programs in C to configuration files for field-programmable gate arrays used in the toolkit for programming of reconfigurable and hybrid computer systems. The creation of parallel program in the COLAMO (Common Oriented Language for Architecture of Multi Objects) language using the analysis results of information dependences in the initial procedural program and its further conversion to a parallel and pipeline form are the distinctive characteristics of the technique. We addressed the methods of scalar splitting and array extension by iterations, which are applied for the fulfillment of the single assignment and unique substitution rules in parallel program and the saving of information communications of the initial procedural program. The technique of conversion of automatically created parallel program to the scalable parallel and pipeline form is presented. The “Procrustes” preprocessor adapts the form for different architectures and configurations of reconfigurable and hybrid computer systems. Owing to the described methodology, it is possible to synthesize a resource-independent scalable COLAMO-application, which can adapt to available computational resource by changing of several constants in automatic mode without any considerable modification of the program source code. Then, the scalable COLAMO-applicationis translated by the COLAMO-translator into field-programmable gate arrays configuration files for the specified reconfigurable computer resource.

TOOLS OF C PROGRAM TRANSLATION FOR RECONFIGURABLE AND HYBRID COMPUTER SYSTEMS BASED ON FIELD-PROGRAMMABLE GATE ARRAYS

In this article, we consider the stages of development of a novel application toolkit for reconfigurable computer systems, its architecture and operation principles. The toolkit provides the translation of procedural programs in C to configuration files for field-programmable gate arrays in 6 stages. The conversion of procedural program in C to parallel program in COLAMO (Common Oriental Language for Architecture of Multi Objects) is performed using four additional programs: the “Angel ” translator and three preprocessors called the “Mermaid ”, the “Procrustes” and the “Nutcracker ”. At the first stage, the “Angel ” translates the C program into the absolutely parallel program in the COLAMO language. During the second stage, the information dependences of the initial C program are analyzed, scalar variables are split, and arrays are stretched by iterations. Then, at the third stage, the COLAMO code is converted to the parallel and pipeline form using the partition of arrays and loops into the vector and strea components. The fourth stage involves the transformation of the COLAMO program to the scalable parallel and pipeline form. The optional fifth stage, which is aimed at the reduction of the base subgraph, is executed only in the case of the lack of hardware resource for the structural implementation of the base subgraph of a problem. At the sixth stage, the COLAMO program is transformed to the unique computational structure and several cadrs designed for the architecture of the chosen reconfigurable computer system. The distinctive feature of the developed toolkit is the support of the resource-independent computing, which allows for the scaling of calculations in both cases of increase (induction) and decrease (reduction) in available hardware resources.

Accuracy of computed tomography: magnetic resonance imaging image fusion using a phantom for skull base surgery.

The aim of this study is to assess the positional accuracy of image fusions of the skull base region using different magnetic resonance imaging (MRI) and computed tomography (CT) image pairs. An image set of 3D fast imaging employing steady-state acquisition-C (FIESTA-C) was used as the base image set. Image fusions were performed using an image set with different fields of view (FOVs): one with different matrix size, one with a different sequence of 3D spoiled gradient recalled acquisition, and one with different modality (CT), using a phantom including multi columnar objects. Position of columns at the center, and 4 and 8 cm from the center were measured. The displacements between the base image set and fused image set were measured. For slices with different z-positions, the displacement of the 8-cm column was assessed. For 20 clinical MRI cases, the distance between the dorsum sellae and the cranial nerves was measured. No significant differences were found between the different FOVs or image sequences. However, with the different matrix sizes and modalities, significant displacements were observed, although they were all within 0.5 mm. Similar displacements were observed in the slices at different z-positions. All cranial nerves were located within 40 mm of the dorsum sellae. The displacements following image fusion were within approximately 0.5 mm, even at 8 cm from the center. This suggests that the region where the cranial nerves are located, within 40 mm of the dorsum sellae, had no risk of positional error following image fusion.

Networked robot identification technology for multi objects based on RFID-ZigBee

In order to identify and monitor a large scale of indoor personnel, this paper combines network robots, RFID and ZigBee, and designs a network robot identification system. CC1110 is used to formulate RFID module, CSMA are combined with TDMA to fulfil multi-card function and overcome data collision. With the processing core S3C6410 and software technology like Linux multi-thread and H.264 decoding, the robot serves as a mobile node to identify IDs and displays guidance video according to targets priority. Gateway module is designed to collect IDs and delivery commands. PC monitoring centre based on Qt can receives IDs and displays them. Test results show that the system can identify targets accurately and provide high-priority goals with guiding service, meets the application requirement.

On-line video multi-object segmentation based on skeleton model and occlusion detection

In this work, we propose an approach for on-line video multi object segmentation based on skeleton model and occlusion detection. We consider the multi-object segmentation in every frame as a multi-class region merging based object segmentation. We then generate the initial object superpixels automatically using a skeleton model from the second frame. Moreover, we also propose an initial background superpixel prediction scheme. In case the occlusion to affect the final segmentation result, we propose an occlusion detection model based on optical flow. The experimental results show that our method is both robust in segmenting multi objects and efficient in execution time.

Visual Persons Behavior Diary Generation Model based on Trajectories and Pose Estimation

The behavior pattern of persons was the important output of the surveillance analysis. This paper focus on the generation model of visual person behavior diary. The pipeline includes the person detection, tracking, and the person behavior classify. This paper adopts the deep convolutional neural model YOLO (You Only Look Once)V2 for person detection module. Multi person tracking was based on the detection framework. The Hungarian assignment algorithm was used to the matching. The person appearance model was integrated by HSV color model and Hash code model. The person object motion was estimated by the Kalman Filter. The multi objects were matching with exist tracklets through the appearance and motion location distance by the Hungarian assignment method. A long continuous trajectory for one person was get by the spatial-temporal continual linking algorithm. And the face recognition information was used to identify the trajectory. The trajectories with identification information can be used to generate the visual diary of person behavior based on the scene context information and person action estimation. The relevant modules are tested in public data sets and our own capture video sets. The test results show that the method can be used to generate the visual person behavior pattern diary with certain accuracy.

Topological optimization method in large-scale objects projects reengineering

She found further development of the mathematical model of the task of reengineering topological structures of large centralized facilities in terms of costs in consideration of multires. The objective function models detailed to the level of explicit parameters depending on the topological structure of the object, which takes into account the cost of upgrading and dismantling of its components. Based on analysis of the characteristics of the objective function proposed modification of directional sorting options using different procedures for determining placement units. As the base used combinatorial procedure that implements exhaustive search all possible variants of the units. To reduce the time complexity of solving the problem of reengineering Topological centralized large scale multi objects in terms of cost, based on the procedures for placing units coordinate-wise optimization methods, simulated annealing and search the prohibitions. They allow for decoupling problem for systems with a large number of structural elements.

Analysis and Detection of Multi Tumor from MRI of Brain using Advance Adaptive Feature Fuzzy C-means (AAFFCM) Algorithm

Objectives: The objective of the study focused on the detection of the multi-tumor must involves evaluation of the computer-aided diagnosis systems which use image processing as the main tool for detection, therefore, the performance parameters that agree with the inter observers must be used. Methods: Segmentation is a significant feature of medical image dispensation, where Clustering move toward is extensively used in biomedical application mainly for brain multi tumor detection in irregular Magnetic Resonance Images (MRI). The present approach derives an innovative method for brain tumor analysis and detection based on the support vector machine (SVM) and fuzzy c-means algorithms.. No such study is available to detect the multi-tumor. The present approach is to solve that problem and used to detect multi-tumors. Findings: The proposed AAFFCM approach is a hybrid approach which is a combination of fuzzy c-means and SVM algorithms for detecting multi-tumors in brain. A color base segmentation technique so as to uses the k-means clustering system is to path the multi tumor objects in the Magnetic Resonance (MR) brain images. Improvement: In the proposed approach, the MRI is improved by improvement techniques such as difference development, and Mean stretch. The skull striping operation is performed by using Morphology and double-thresh holding technique. By using Matrix, the specific information is removed from the brain image which is called Grey Level Advance Length Matrix (GLALM). After removing the specific information from the brain, SVM algorithm is used to categorize the brain MRI images, which give precise and more effectual importance for categorization of brain MRI.

Reliability analysis of RC T-beam highway bridges in China based on a virtual bridge dataset

Limit state design method based on probability theory is used in the Chinese highway bridge design codes which adopt partial factors in the design equations. In this paper, a virtual bridge dataset (VBD) including a total of 8064 RC T-beam bridges designed based on the Chinese highway bridge design codes was established. With this VBD, the influences of the different design parameters including span length, girder spacing, concrete grade, reinforcement grade, number of design lanes, and actions on the reliability of the RC T-beams were studied. A Monte-Carlo simulation combined with first order second moment method was employed to calculate the reliability indices of the bridges in the VBD. By conducting a statistical analysis on the calculated reliability indices, it is shown that the reliability indices follow a normal distribution and there is good consistency in the reliability of the bridges with different parameters. Construction cost for the bridges with spans of 16m in the VBD was calculated, indicating that the construction cost increases along with the increase of the reliability indices. Suggestions with regard to the determination of the target reliability of the bridges in multi objects were also proposed.

Implementation of Object Recognition Based on Local Texture Operators in FPGA Technology

In the arrival of today’s highly integrated multimedia device and fast emerging applications, image processing have become more important than any others. These devices require complex image processing tasks lead to a very challenge design process as it demands more efficient and high processing systems. The scope of the project is to extract locations and features of multi objects in an image for object recognition. For low power consumption and better performance we design a proposed system in FPGA. In existing system an cell based multi object feature extraction algorithm is used to extract simultaneously autocorrelation feature of objects. It is calculated using zeroth order and first order moments to obtain the size and location of multiple objects. To reduce computational complexity and memory consumption, Local Binary Pattern (LBP) and Local Ternary Pattern (LTP) is used to extract the feature of multiple objects are proposed. In the local binary pattern, the LBP value is computed by comparing a gray level value of centre pixel in an image with its neighbors. The local ternary pattern is extended from LBP to three-valued code in which gray values are quantized to zero,+1,-1. The proposed architecture is designed using verilog HDL, simulated using Modelsim software and synthesized using Xilinx project navigator.

Foreground Object Extraction Based on Interactive Color Saliency Map

Efficient interactive foreground extraction from color images is of great practical importance in computer vision. In recent years, an approach based on optimization using graph-cuts has been widely used. However, an interactive foreground extraction approach, which is intuitive and efficient for users, is required for practical applications. In this paper, we propose a novel foreground object extraction method based on interactive color saliency map. In our method, the user provides only some foreground pixels as the initial reference. To achieve extraction of multi target objects, likelihoods of the foreground and background are defined by a color saliency map based on the provided reference pixels. These likelihoods are added to the cost of the proposed graph. Finally, image segmentation is performed by optimizing the proposed graph using the graph-cut algorithm.

Hierarchical Intelligent Alignment Control of Air Spring Vibration Isolation System of Ship Propulsion Plant

Alignment control of propulsion plant air spring vibration isolation system is a complicated Multi objects control problem. In this paper a new hierarchical intelligent control method of alignment control is brought out. Known from current hierarchical intelligent control methods with three layers, this method consists of task layout layer, control decision-making layer, control layout layer and control executive layer. The complicated alignment control task is decomposed into four hierarchical layers to reduce complexity. The control system has good control convergence capability and fast convergence speed. Test results validate the feasibility of this method.

Interactive Segmentation for Color Image based on Color Saliency

Efficient interactive foreground extraction from color images is of great practical importance in computer vision. In recent yaers, an approach, based on optimization using graph-cuts has been widely used. However, an interactive foreground extraction approach, which is intuitive and efficient for users, is required for practical applications. In this paper, we propose a novel interactive foreground extraction method based on color saliency. In our method, the user provides only some foreground pixels as the initial reference. To achieve extraction of multi target objects, likelihoods of the foreground and background are defined by a gaussian mixture model and a color saliency map based on the provided reference pixels. These likelihoods are added to the cost of the proposed graph. Finally, image segmentation is performed by optimizing the proposed graph using the graph-cut algorithm.

The Shock Isolation System Optimum Design Based on Pseudo Parallel Improved Genetic Algorithm

Based on the vertical maximum displacement response, the paper uses a new pseudo—parallel improved genetic algorithm (PPIGA) to optimize mass and stiffness of the shock isolation system. In the process of optimization the Newmark procedure is used to calculate shock response. The optimization results show that the optimization parameter (mass and stiffness) can satisfy restricting condition and the multi—objects optimization can be realized．