![]() The PSO delivers good average performances and is relatively easier to implement. PSO, unlike PF, allows particles to interact with one another, and the interactions has been shown to be effective in finding global optima in high-dimensional search space. Particle Swarm Optimization (PSO), in particular, has been becoming popular in human motion tracking. These methods have the ability to approximate highly non-linear problem, with relatively robust and reliable performance, and with relatively fewer tuning parameters. ![]() More recently, stochastic global optimization methods such as the population-based evolutionary algorithms (EAs) and swarm intelligence (SI) have been gaining popularity. Most of the solutions based on local optimization however suffer from the curse of dimensionality and rely on simple human models (which lead to suboptimal tracking results) or require a high number of evaluations to provide satisfactory results. To tackle the high-dimensionality of the problem, some solutions partition the search space while some others utilize multiple stage search operation. Until recently, most recent work are based on variants of local optimization method such as particle filtering (PF). Many solutions have been proposed for model-based articulated human motion tracking. Other challenges include cluttered background, occlusion, ambiguity and illumination changes. The key challenge in the approach is the high-dimensionality of the search space involved, due to the large number of degrees of freedom (DOF) typically present in an articulated human body figure. The optimization problem then becomes that of determining the body model configuration which will result in the best match to the images in the video. The key idea is to render the body model and to compare the rendered image with acquired video frame in order to determine the fitness of the body model configuration. A dominant line of approaches to the task is one that utilizes a 3D articulated human body model. The primary objective of markerless articulated human motion tracking is to automatically localize the pose and position of a subject from a video stream (sequences of images). Markerless articulated human motion tracking is an emerging field with potential applications in areas such as automatic smart security surveillance, medical rehabilitation, and 3D animation industries. Comprehensive experimental results are presented to support the claims. Further, the proposed tracking method is capable of automatic initialization and self-recovery from temporary tracking failures. Experiments using Brown and HumanEva-II dataset demonstrated that H-MCPSO performance is better than two leading alternative approaches-Annealed Particle Filter (APF) and Hierarchical Particle Swarm Optimization (HPSO). Both the silhouette and edge likelihoods are used in the fitness function. The tracking problem is formulated as a non-linear 34-dimensional function optimization problem where the fitness function quantifies the difference between the observed image and a projection of the model configuration. To overcome these drawbacks, we have developed a method for the problem based on Hierarchical Multi-Swarm Cooperative Particle Swarm Optimization (H-MCPSO). However, the classical PSO suffers from premature convergence and it is trapped easily into local optima, significantly affecting the tracking accuracy. The high-dimensional search space involved in markerless full-body articulated human motion tracking from multiple-views video sequences has led to a number of solutions based on metaheuristics, the most recent form of which is Particle Swarm Optimization (PSO).
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |