JCP 2013 Vol.8(10): 2478-2482 ISSN: 1796-203X
doi: 10.4304/jcp.8.10.2478-2482
doi: 10.4304/jcp.8.10.2478-2482
A Method of Target Region Detection Based on Multi-channel Weighted Visual Attention
Hui Wang1, Yuanyuan Dang1, Hongchang Ke2, and Xiaonan Liu3
1 College of Computer Science and Engineering, Changchun University of Technology, Changchun, China
2 Software Vocational Institute of Technology, Changchun Institute of Technology, Changchun, China
3 China Information Technology Security Evaluation Center, Beijing, China
Abstract—For traditional imagery attention model, the extraction of the bottom-up visual attention information of each channel is not thinking about the disadvantages of topdown task. A top-down guidance based feature vector weighted integration strategy is designed and implemented to compare the multi-feature-channel features of the target and the whole image. Based on the strategy, the weight of each feature channel in integration and the different map between the target and each image can be calculated to guide the generation of saliency map. A new method of target region detection based on multi-channel weighted visual attention is proposed to provide a calculation mechanism for quick location on images. The experimental results indicate that the method proposed in paper can search and lock the region more quickly which the target is most likely exist, and improve the efficiency of target searching.
Index Terms—visual attention, top-down guidance, different map, multi-channel weighted
2 Software Vocational Institute of Technology, Changchun Institute of Technology, Changchun, China
3 China Information Technology Security Evaluation Center, Beijing, China
Abstract—For traditional imagery attention model, the extraction of the bottom-up visual attention information of each channel is not thinking about the disadvantages of topdown task. A top-down guidance based feature vector weighted integration strategy is designed and implemented to compare the multi-feature-channel features of the target and the whole image. Based on the strategy, the weight of each feature channel in integration and the different map between the target and each image can be calculated to guide the generation of saliency map. A new method of target region detection based on multi-channel weighted visual attention is proposed to provide a calculation mechanism for quick location on images. The experimental results indicate that the method proposed in paper can search and lock the region more quickly which the target is most likely exist, and improve the efficiency of target searching.
Index Terms—visual attention, top-down guidance, different map, multi-channel weighted
Cite: Hui Wang, Yuanyuan Dang, Hongchang Ke, and Xiaonan Liu, " A Method of Target Region Detection Based on Multi-channel Weighted Visual Attention," Journal of Computers vol. 8, no. 10, pp. 2478-2482, 2013.
General Information
ISSN: 1796-203X
Abbreviated Title: J.Comput.
Frequency: Bimonthly
Abbreviated Title: J.Comput.
Frequency: Bimonthly
Editor-in-Chief: Prof. Liansheng Tan
Executive Editor: Ms. Nina Lee
Abstracting/ Indexing: DBLP, EBSCO, ProQuest, INSPEC, ULRICH's Periodicals Directory, WorldCat,etc
E-mail: jcp@iap.org
-
Nov 14, 2019 News!
Vol 14, No 11 has been published with online version [Click]
-
Mar 20, 2020 News!
Vol 15, No 2 has been published with online version [Click]
-
Dec 16, 2019 News!
Vol 14, No 12 has been published with online version [Click]
-
Sep 16, 2019 News!
Vol 14, No 9 has been published with online version [Click]
-
Aug 16, 2019 News!
Vol 14, No 8 has been published with online version [Click]
- Read more>>