원문정보
보안공학연구지원센터(IJMUE)
International Journal of Multimedia and Ubiquitous Engineering
Vol.11 No.4
2016.04
pp.233-242
피인용수 : 0건 (자료제공 : 네이버학술정보)
초록
영어
To deal with rough image region segmentation results, a common method of bag-of-words in statistical text is proposed in the paper. The method splits regional objects to numerous small image blocks, from which rough semantic concepts of regional objects are fetched; then, through the application of multi-instance learning idea and computation of type confidence degree of each rough semantic concept, the impacts of type errors on such concepts can be effectively eliminated, and thus feature semantic concepts of various regional object type are obtained.
목차
Abstract
1. Introduction
2. Classifying Method of Multi-Instance Image Regions
3. Extraction of Rough Semantic Concepts
4. Extraction of Feature Semantic Concepts
5. Experiment Design and Discussion
5.1. Training Stage
5.2. Testing Stage
6. Conclusion
References
1. Introduction
2. Classifying Method of Multi-Instance Image Regions
3. Extraction of Rough Semantic Concepts
4. Extraction of Feature Semantic Concepts
5. Experiment Design and Discussion
5.1. Training Stage
5.2. Testing Stage
6. Conclusion
References
저자정보
참고문헌
자료제공 : 네이버학술정보