Human and car identification using motion vector in H.264 compressed video

Author

Chen, Wei ; Yang, Quan-Xi ; Lin, Ke-Wei ; Wang, Sheng-Yu ; Huang, Chung-Lin

Author_Institution

Dept. of Electr. Eng., Nat. Tsing Hua Univ., Hsinchu, Taiwan

fYear

2011

fDate

6-9 Nov. 2011

Firstpage

1

Lastpage

4

Abstract

This paper presents a novice method for human and car identification in H.264/AVC compressed video domain. By analyzing the shape and motion vector homogeneity of the segmented objects, we can identify car and human. Our system consists of three main processes: (1) Moving object segmentation based on clustering MVs and Markov Random Field (MRF) iteration, (2) Feature Extraction based on motion analysis to obtain the difference of MVs direction (dMVD) and shape analysis to find the number of MBs (nMB) of an object, and (3) Object classification using Bayesian Classifier. In the experiments, we show that the recognition rate of car and human are 88% and 98% respectively.

Keywords

Bayes methods; Markov processes; automobiles; feature extraction; image motion analysis; image recognition; image segmentation; vectors; video coding; video surveillance; Bayesian classifier; H.264 compressed video; MRF iteration; Markov random field; car identification; feature extraction; human identification; motion analysis; motion vector; moving object segmentation; object classification; shape analysis; shape vector; Algorithm design and analysis; Humans; Motion segmentation; Object recognition; Object segmentation; Shape; Streaming media;

fLanguage

English

Publisher

ieee

Conference_Titel

Visual Communications and Image Processing (VCIP), 2011 IEEE

Conference_Location

Tainan

Print_ISBN

978-1-4577-1321-7

Electronic_ISBN

978-1-4577-1320-0

Type

conf

DOI

10.1109/VCIP.2011.6115985

Filename

6115985