Michael Kamaraj - Academia.edu (original) (raw)

Uploads

Papers by Michael Kamaraj

Research paper thumbnail of Optimization of Multi-Target Tracking and Occlusion Handling Using Mean Shift Method

Multi target tracking is an interesting and challenging task in finding an optimal set of path wi... more Multi target tracking is an interesting and challenging task in finding an optimal set of path within a temporal window. The problem of multi target tracking comprises of few distinct challenges, the naturally discrete problem of data association, and continuous problem of trajectory estimation. Many recent approaches often perform multi-target tracking as discrete optimization which need a pre-computation and time. Alternatively, a framework is designed to focus on complete representation of the problem. In this work, an energy term is formulated as minimization of continuous energy in multiple target tracking. The energy function includes the dynamic model of target, mutual exclusion, track persistence and regularization. In addition, the occlusion and ambiguous targets of appearances are handled using the mean shift clustering which is largely invariant to both partial and full occlusions, complex backgrounds and change in scale. The mean shift introduces a feature space analysis...

Research paper thumbnail of Human Motion tracking using Gaussian Mixture Method and Beta-Likelihood Matching

Video surveillance is widely used to monitor the place which needs constant security such as Bank... more Video surveillance is widely used to monitor the place which needs constant security such as Banks, Shopping Malls, Highways, crowded public places, country borders etc. The major disputes include the complex motion behaviours of different human objects, complex scenes with numerous targets, detection of change in human motion. The objective of this paper is to develop a visual detection and tracking system of observing moving objects. We propose the GMM-Likelihood matching Method of tracking algorithm which integrates the adaptive best background detection, data association, adding new hypothesis update kalman measurement, and linear assignment problem to minimise the cost of observation of tracking. The experimental result shows that the active background can be extracted accurately and expeditiously, the algorithm is more robust, and can be utilized in the real time tracking applications. keywords : Real-time visual tracking, Active background estimation, Activity modelling, Data...

Research paper thumbnail of Global Energy Minimization and Optimization of Multi-Target Tracking

Journal of Computational and Theoretical Nanoscience, 2017

The recent multiple target tracking methods aim to obtain the best possible number of trajectorie... more The recent multiple target tracking methods aim to obtain the best possible number of trajectories within the time frame, and few constraints have been set to handle the wide area of trajectories by discrete mapping. In this novel approach of multi target tracking, energy terms are formulated to attain the global optimization which includes the entire representation of the issues such as target tracking, operational representation, collision handling and trajectory processing. Furthermore, two optimization strategies such as the gradient descent which is performed on multiple feature space to obtain local minima of a density function from the given sample of data and gradient ascent which is carried out to achieve a likelihood matching of the target and used to handle the partial evidence of the image, and also uncertainty of the various targets are minimized. The experimentation is performed on the openly available dataset and the mean target tracking accuracy and precision is stud...

Research paper thumbnail of An improved motion detection and tracking of active blob for video surveillance

2013 Fourth International Conference on Computing, Communications and Networking Technologies (ICCCNT), 2013

Motion detection plays a primary vital part in the process of extraction of information regarding... more Motion detection plays a primary vital part in the process of extraction of information regarding the moving objects and utilizes the stabilization in functional areas, such as tracking, classification, recognition etc. In this paper, we propose an extended view and precise approach to motion detection for the automatic video surveillance system. This method accomplishes a complete detection of any moving object by involving substantial modules such as Background subtraction method, Dynamic Optimization threshold method to gain a vivid picture of the moving object, Morphological filtering is introduced to eliminate the noise and hence Background disturbance problem is solved. Then the paper presents tracking and cropping of the detected moving object. The visual inspection analysis of this method of Detecting, Tracking and Cropping of moving object has gained a better result.

Research paper thumbnail of Multiple Target Tracking Using Cost Minimization Techniques

ICTACT Journal on Image and Video Processing, 2017

Research paper thumbnail of Optimization of Multi-Target Tracking and Occlusion Handling Using Mean Shift Method

Multi target tracking is an interesting and challenging task in finding an optimal set of path wi... more Multi target tracking is an interesting and challenging task in finding an optimal set of path within a temporal window. The problem of multi target tracking comprises of few distinct challenges, the naturally discrete problem of data association, and continuous problem of trajectory estimation. Many recent approaches often perform multi-target tracking as discrete optimization which need a pre-computation and time. Alternatively, a framework is designed to focus on complete representation of the problem. In this work, an energy term is formulated as minimization of continuous energy in multiple target tracking. The energy function includes the dynamic model of target, mutual exclusion, track persistence and regularization. In addition, the occlusion and ambiguous targets of appearances are handled using the mean shift clustering which is largely invariant to both partial and full occlusions, complex backgrounds and change in scale. The mean shift introduces a feature space analysis...

Research paper thumbnail of Human Motion tracking using Gaussian Mixture Method and Beta-Likelihood Matching

Video surveillance is widely used to monitor the place which needs constant security such as Bank... more Video surveillance is widely used to monitor the place which needs constant security such as Banks, Shopping Malls, Highways, crowded public places, country borders etc. The major disputes include the complex motion behaviours of different human objects, complex scenes with numerous targets, detection of change in human motion. The objective of this paper is to develop a visual detection and tracking system of observing moving objects. We propose the GMM-Likelihood matching Method of tracking algorithm which integrates the adaptive best background detection, data association, adding new hypothesis update kalman measurement, and linear assignment problem to minimise the cost of observation of tracking. The experimental result shows that the active background can be extracted accurately and expeditiously, the algorithm is more robust, and can be utilized in the real time tracking applications. keywords : Real-time visual tracking, Active background estimation, Activity modelling, Data...

Research paper thumbnail of Global Energy Minimization and Optimization of Multi-Target Tracking

Journal of Computational and Theoretical Nanoscience, 2017

The recent multiple target tracking methods aim to obtain the best possible number of trajectorie... more The recent multiple target tracking methods aim to obtain the best possible number of trajectories within the time frame, and few constraints have been set to handle the wide area of trajectories by discrete mapping. In this novel approach of multi target tracking, energy terms are formulated to attain the global optimization which includes the entire representation of the issues such as target tracking, operational representation, collision handling and trajectory processing. Furthermore, two optimization strategies such as the gradient descent which is performed on multiple feature space to obtain local minima of a density function from the given sample of data and gradient ascent which is carried out to achieve a likelihood matching of the target and used to handle the partial evidence of the image, and also uncertainty of the various targets are minimized. The experimentation is performed on the openly available dataset and the mean target tracking accuracy and precision is stud...

Research paper thumbnail of An improved motion detection and tracking of active blob for video surveillance

2013 Fourth International Conference on Computing, Communications and Networking Technologies (ICCCNT), 2013

Motion detection plays a primary vital part in the process of extraction of information regarding... more Motion detection plays a primary vital part in the process of extraction of information regarding the moving objects and utilizes the stabilization in functional areas, such as tracking, classification, recognition etc. In this paper, we propose an extended view and precise approach to motion detection for the automatic video surveillance system. This method accomplishes a complete detection of any moving object by involving substantial modules such as Background subtraction method, Dynamic Optimization threshold method to gain a vivid picture of the moving object, Morphological filtering is introduced to eliminate the noise and hence Background disturbance problem is solved. Then the paper presents tracking and cropping of the detected moving object. The visual inspection analysis of this method of Detecting, Tracking and Cropping of moving object has gained a better result.

Research paper thumbnail of Multiple Target Tracking Using Cost Minimization Techniques

ICTACT Journal on Image and Video Processing, 2017