Title: a shot boundary detection method based on context vector and tabu-svm
Author: Zhao Long
; Sun Xuemei
; Shang Bingjian
; Wu Jigang
Keyword: Algorithms
; Image retrieval
; Image segmentation
; Tabu search
; Video signal processing
Source: Journal of Computational Information Systems
Issued Date: 2012
Volume: 8, Issue: 16, Pages: 6817-6824 Indexed Type: EI
Department: (1) School of Computer Science and Software Engineering Tianjin Polytechnic University Tianjin 300387 China; (2) State Key Laboratory of Computer Science Institute of Software Chinese Academy of Sciences Beijing 100190 China
Abstract: Categorizing the consecutive video frames into shots is the first and prerequisite step for contentbased video indexing and retrieval. With a considerable amount of research focusing on the shot boundary detection, traditional classification for common shot transitions has achieved satisfactory results. However, most of the existing algorithms are susceptible to the effect from thresholds, and can't avoid missing the long gradual shot transitions hard to detect. To address the two problems, a novel method which uses context feature vector and Tabu-SVM is presented in the paper. The major advantages of the proposed algorithm are quite accurate detection performance and optimization towards the classification model. Experimental results show that our proposed method is effective and robust, especially the boundaries of long transitions are detected with minimal inaccuracy. © 2012 Binary Information Press. English Abstract: Categorizing the consecutive video frames into shots is the first and prerequisite step for contentbased video indexing and retrieval. With a considerable amount of research focusing on the shot boundary detection, traditional classification for common shot transitions has achieved satisfactory results. However, most of the existing algorithms are susceptible to the effect from thresholds, and can't avoid missing the long gradual shot transitions hard to detect. To address the two problems, a novel method which uses context feature vector and Tabu-SVM is presented in the paper. The major advantages of the proposed algorithm are quite accurate detection performance and optimization towards the classification model. Experimental results show that our proposed method is effective and robust, especially the boundaries of long transitions are detected with minimal inaccuracy. © 2012 Binary Information Press. Language: 英语
Content Type: 期刊论文
URI: http://ir.iscas.ac.cn/handle/311060/15037
Appears in Collections: 软件所图书馆_期刊论文
There are no files associated with this item.
Recommended Citation:
Zhao Long,Sun Xuemei,Shang Bingjian,et al. a shot boundary detection method based on context vector and tabu-svm[J]. Journal of Computational Information Systems,2012-01-01,8(16):6817-6824.
