An Online Learning Approach to Occlusion Boundary Detection

NANO SCIENTIFIC RESEARCH CENTRE PVT.LTD.,  AMEERPET, HYD

WWW.NSRCNANO.COM, 09640648777, 09652926926

  

JAVA PROJECTS LIST--2013

JAVA 2013 IEEE PAPERS

 An Online Learning Approach to Occlusion Boundary Detection

Abstract:

            We propose a novel online learning-based framework for occlusion boundary detection in video sequences. This approach does not require any prior training and instead “learns” occlusion boundaries by updating a set of weights for the online learning Hedge algorithm at each frame instance. Whereas previous training-based methods perform well only on data similar to the trained examples, the proposed method is well suited for any video sequence. We demonstrate the performance of the proposed detector both for the CMU data set, which includes hand-labeled occlusion boundaries, and for a novel video sequence. In addition to occlusion boundary detection, the proposed algorithm is capable of classifying occlusion boundaries by angle and by whether the occluding object is covering or uncovering the background.

Existing System:

            The presence of occlusions in the image and video processing literature is astoundingly diverse. Research is conducted both to determine occlusion boundaries explicitly and to use this information implicitly in the pursuit of other results. Implicit determination of occlusion boundaries has been cited in applications of object tracking, segmentation, and disparity estimation.

Proposed System:

            The online learning framework is based on the idea of a panel of experts, where each expert is proficient at a distinct task. Each time a decision needs to be made, each expert is polled, and a loss is calculated based on how correct their prediction turned out to be. Over time, the best expert for each task will become evident. In this paper, the task of each expert is to detect a certain type of occlusion event, and the quality of each expert’s detection is measured using a loss function in the pixel domain (similar to a standard distortion measure). At a specific location, an occlusion boundary is detected based on the best prediction over all experts. Contrary to all competing approaches, the proposed method does not require any training; rather, the classification of occlusion boundaries is based on the relative weighting of the experts, which occurs online.    

       

Software and Hardware Requirements

Hardware Required:                            

System                                    :           Pentium IV

Hard Disk                   :           80 GB

RAM                           :           512 MB

Software Required:

Operating System       :           Windows XP

Language                    :           java

Modules:

·         Access Video Source

·         Boundary Detection

font-siH � 2 0 �� �-� eight:115%;font-family:"Times New Roman","serif"; mso-fareast-font-family:Calibri;mso-ansi-language:EN-US;mso-fareast-language: EN-US;mso-bidi-language:AR-SA'>Trapdoor Hash Verification