Abstrakt

Particle Filtering Framework and Occlusion Handling Using Region Based Tracking Technique for 2-D and 3-D Image Sequences

J.Cynthia Esther Rani, R.Sathya, R.D. Sathiya

A method is introduced to address the estimation of pose for 2D-3D images. This approach involves the tracking of 2D images in 3D spaces which results in the estimation of pose with the continuous process of particle filtering and handling the scheme in the presence of occlusion. In order to maintain the track and to detect the exact region, a unique method Region Based Tracking method is used in this paper. It also involves the degree of dependencies between predictions and measurements model which determines the position and estimation of the object sustaining particle filters and handling occlusions. Particle filtering can propagate via autoregressive model for tracking 2D images but also estimating a 3D pose. Finally, this method is considered to be more effective in both robustness and speed compared to similar video tracking and pose estimation methods. Thus the conversion of pose estimation from 2D Silhouette curve to 3D Occluding Curve is finally illuminated in video.