Title :
A Moving Object Segmentation in MPEG Compressed Domain Based on Motion Vectors and DCT Coefficients
Author :
Liu, Long ; Feng, Xingle ; Ji, Ruirui ; Deng, Yi
Abstract :
More and more attention has been paid to the study of moving object segmentation in MPEG compressed domain these years. In this paper, a straightforward scheme by using motion vectors and DCT coefficients is put forward for segmenting moving object in MPEG compressed domain. The feature vector of DCT is defined. As regards local motion scene, the scheme produces a better result of segmentation than 1currently adopted ones. Firstly, the motion vectors are processed by accumulation and median filter; then the areas of non-zero motion vectors are rectified by correlations of DCT blocks to get high-quality segmentation.
Keywords :
Clustering algorithms; Discrete cosine transforms; Face detection; Filters; Flowcharts; Layout; Object segmentation; Signal processing algorithms; Transform coding; Video compression;
Conference_Titel :
Image and Signal Processing, 2008. CISP '08. Congress on
Conference_Location :
Sanya, China
Print_ISBN :
978-0-7695-3119-9
DOI :
10.1109/CISP.2008.7