Approach to model human appearance based on sparse representation for human tracking in surveillance

Damotharasamy, Sangeetha

doi:10.1049/iet-ipr.2018.5961

Cited by 10 publications

(15 citation statements)

References 66 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

Section: Introductionmentioning

confidence: 99%

“…The most significant task assigned in the domain of computer vision is the human tracking, which aim to recognize the location of the human at any edge of the world (Damotharasamy, 2020). Home land security (Chowdhry, et al, 2015) , gait characterization (Ran, et al, 2010), gender classification (Chen & Hsieh, 2012) , crowd flow analysis (Eshel & Moses, 2008),accident prediction, crime prevention, close-packed crowd (Lai, et al, 2013), fall detection of elderly (Thome, et al, 2008),driver assistance (Akhlaq, et al, 2012), smart video surveillance, monitoring the children at home (Liu, et al, 2017), monitoring the patients and to monitor the human activities (Reddy & Shah, 2013) are some of the application of the human tracking (Damotharasamy, 2020) .Yet detecting the presence of humans in the surveillance video is the hectic challenge in computer vision due to the various constraints, such as the variation in the human poses,the view point of the camera, illumination and occlusion. The deviation in exterior appearance of human is experienced with respect to their poses along with the variation in stand point as the human anatomy is comprised of various joints (Dilawari, et al, 2020).The camera should be fixed such that the view point not exceeds the angle of 45• to find out the human beings at any orientation yet it can recognize the human in a moderately upstanding posture.Furthermore, the human shape evaluation demands sufficient resolution and the width of the human image areneeded to be more than 24px.…”

Section: Introductionmentioning

confidence: 99%

“…The human tracking algorithm is characterized into two significant methodologies, specifically known as discriminative and the generative methodologies. The most analogues regions in the image are scrutinized in order to recognize the target model in the Generative methodology (Damotharasamy, 2020). The fundamental idea is to accomplish a representation that can precisely portray the framework, and then to make a deduction between the current edge picture and the foundation model, and afterward select an edge (Zhang, 2019).…”

Section: Introductionmentioning

confidence: 99%

“…The whole-based tracking strategy shows prevalent execution when the article is generally furnished in the scene. However, it experiences some impudent in the presence of the occlusion (Damotharasamy, 2020).Hence, the recognition execution of the subsequent models isn't acceptable in complex scenes, while gathering adequate and exacting comments through bouncing boxes in huge scope and different situations is complex and requires high cost. The generative strategy utilizes the classifiers to differentiate the closer view objects from the background sectors which are developed to suppress the drawbacks of the generative methods.…”

Section: Introductionmentioning

confidence: 99%

“…The generative strategy utilizes the classifiers to differentiate the closer view objects from the background sectors which are developed to suppress the drawbacks of the generative methods. Now-a-days various classification algorithms are developed to track the human in the videos or images (Damotharasamy, 2020). The researchers utilized the background subtraction method (Zhang, et al, 2020) in order to update the quantity of the models, learning rates and the weights.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Hybrid Tracking Model Based on Optimized Deep Learning for Object Detection and Tracking Using the Multi-View Cameras

Dewarde¹,

Kotrappa²

2021

INDJCSE

View full text Add to dashboard Cite

Video surveillance renders great services to humans formonitoring the children, crime prevention and monitoring the patients. Yet, detecting the presence of human objects in the surveillance video is found to be hectic challenge due to various constraints, such as the view point of the camera, illumination and occlusion. In this research, a hybrid model is proposed to restrain the issues, like illumination and occlusions for effective recognition of human objects in the surveillance video. In this proposed hybrid model, the object in the frame are tracked with the aid of the rectangular bounding box, proposed search source-based Deep LSTM, and modified deep sort algorithm, where the effect of vanishing gradient is minimized and performance is boosted. The search space algorithm trains the Deep LSTM and establishes an effective trade-off between the exploration and exploitation phases with a better hyperparameters that impacts directly on the tracking performance. At the same time, the modified deep sort algorithm solves the assignment problem through incorporating the motion information, which in turn increases the tracking efficiency. The performance evaluation and comparative analysis are executed to demonstrate the effectiveness of the proposed Hybrid model. The proposed Hybrid model shows impressive results in terms of MOTA, MOTP and the tracking error of98.5%, 98.1% and 0.0135 respectively.

show abstract

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Hybrid Tracking Model Based on Optimized Deep Learning for Object Detection and Tracking Using the Multi-View Cameras

Dewarde¹,

Kotrappa²

2021

INDJCSE

View full text Add to dashboard Cite

show abstract

Application of improved virtual sample and sparse representation in face recognition

Zhang

Wang

Zhang

et al. 2022

CAAI Trans on Intel Tech

View full text Add to dashboard Cite

Sparse representation plays an important role in the research of face recognition. As a deformable sample classification task, face recognition is often used to test the performance of classification algorithms. In face recognition, differences in expression, angle, posture, and lighting conditions have become key factors that affect recognition accuracy. Essentially, there may be significant differences between different image samples of the same face, which makes image classification very difficult. Therefore, how to build a robust virtual image representation becomes a vital issue. To solve the above problems, this paper proposes a novel image classification algorithm. First, to better retain the global features and contour information of the original sample, the algorithm uses an improved non‐linear image representation method to highlight the low‐intensity and high‐intensity pixels of the original training sample, thus generating a virtual sample. Second, by the principle of sparse representation, the linear expression coefficients of the original sample and the virtual sample can be calculated, respectively. After obtaining these two types of coefficients, calculate the distances between the original sample and the test sample and the distance between the virtual sample and the test sample. These two distances are converted into distance scores. Finally, a simple and effective weight fusion scheme is adopted to fuse the classification scores of the original image and the virtual image. The fused score will determine the final classification result. The experimental results show that the proposed method outperforms other typical sparse representation classification methods.

show abstract

Low‐rank nonnegative sparse representation and local preservation‐based matrix regression for supervised image feature selection

Zhu

Chen

2021

IET Image Processing

View full text Add to dashboard Cite

Matrix regression has attracted much attention due to directly select some meaningful features from matrix data. However, most existing matrix regressions do not consider the global and local structure of the matrix data simultaneously. To this end, we propose a lowrank nonnegative sparse representation and local preserving matrix regression (LNSRLP-MR) model for image feature selection. Here, the loss function is defined by the left and right regression matrices. To capture the global structure and discriminative information of the training images and reduce the effect of heterogeneous data and noises, we impose the low-rank constraint on the self-representation error matrix and the nonnegative sparse constraint on the coefficient vector. The graph matrix can be learned adaptively through representation coefficients, so that accurate local structure information in samples can be revealed. Feature selection is performed by obtained row sparse transformation matrix. An optimization procedure and its performance are also present. Experimental results on several image datasets show that compared with the-state-of-the-art method, the average classification accuracy of the proposed method is improved by at least 1.2% and up to 3.3%. For images with noise or occlusion, the accuracy is improved significantly, up to 4%, which indicates that this method has strong robustness.This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.

show abstract

Approach to model human appearance based on sparse representation for human tracking in surveillance

Cited by 10 publications

References 66 publications

Hybrid Tracking Model Based on Optimized Deep Learning for Object Detection and Tracking Using the Multi-View Cameras

Hybrid Tracking Model Based on Optimized Deep Learning for Object Detection and Tracking Using the Multi-View Cameras

Application of improved virtual sample and sparse representation in face recognition

Low‐rank nonnegative sparse representation and local preservation‐based matrix regression for supervised image feature selection

Contact Info

Product

Resources

About