원문정보
보안공학연구지원센터(IJMUE)
International Journal of Multimedia and Ubiquitous Engineering
Vol.10 No.6
2015.06
pp.357-364
피인용수 : 0건 (자료제공 : 네이버학술정보)
초록
영어
In this paper, we proposed an algorithm which efficiently detects salient objects. First of all, yellow and intensity images are generated from input image and the feature map is generated using a contrast difference between red-green and blue-yellow channel. Secondly, Lookup-Table(LUT) is obtained by histogram of the feature map. Finally, the saliency map is obtained from LUT. The proposed saliency detector shows that this method is able to efficiently extract saliency maps by improving the AUC 10-40%, comparing to existing color-based method and frequency space transform method.
목차
Abstract
1. Introduction
2. Related Works
2.1 IT Method
2.2 Spectral Residual (SR) Method
3. Proposed Algorithm
3.1 Generation of Yellow and Intensity Maps
3.2 Extraction of Color Feature Images
3.3 Generation of Histogram and LUT
3.4 Conversion of Histograms to a Saliency Map
4. Experiment and Analysis
4.1 Experiment
4.2 Analysis
5. Conclusions
Acknowledgements
References
1. Introduction
2. Related Works
2.1 IT Method
2.2 Spectral Residual (SR) Method
3. Proposed Algorithm
3.1 Generation of Yellow and Intensity Maps
3.2 Extraction of Color Feature Images
3.3 Generation of Histogram and LUT
3.4 Conversion of Histograms to a Saliency Map
4. Experiment and Analysis
4.1 Experiment
4.2 Analysis
5. Conclusions
Acknowledgements
References
저자정보
참고문헌
자료제공 : 네이버학술정보