Senin, 21 Juli 2008

Temporal Coherence Extraction based on Motion Vector

Abstract--This final project present a method to extract temporal coherence on video using motion vector theory. At recent days, there are so many applications using temporal coherence as their based, such as video compressing applications [1], objects movement prediction applications [2], bounding volume hierarchies for collision detection application [3]. Temporal coherence is an assumption that objects on a video sequences are moves smoothly and from that assumption we can get that between two sequencing frames have connections or coherences. In this project, the writer using motion vector as base method to extract the temporal coherence. By analyzing the motion vectors, we can get the coherences among between two sequencing frame. This project use block matching algorithm as base algorithm to find motion vectors between two frames. From the experiment, we will know how sampling values should used on the extraction process, what factors may fail the extraction, and the final goal is, we can get the temporal coherence of a video sequence.

    Keywords :

    Ekstraksi, temporal coherence, video, motion vector, block matching, POC.


    Download the article :