15 November 2011 Image classification by multi-instance learning with base sample selection
Author Affiliations +
Proceedings Volume 8335, 2012 International Workshop on Image Processing and Optical Engineering; 833519 (2011) https://doi.org/10.1117/12.917409
Event: 2012 International Workshop on Image Processing and Optical Engineering, 2012, Harbin, China
Abstract
We propose a similarity-based learning style algorithm by regarding each image as a multi-instance (MI) sample for image classification. An image featured as vectorial representation interesting regions is transferred to a MI sample. Then a similarity like matrix is constructed using MI kernel between given images and some carefully selected base images, as the new representation of given images. Three selection strategies are proposed to build the base images set to find an optimal solution. A Weka implementation decision tree is used as the main learner in this paper. Experiments on image data repository ALOI and Corel Image 2000 show the effectiveness of the proposed algorithm compared to some previous based line methods.
© (2011) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Qiang Pan, Qiang Pan, Gang Zhang, Gang Zhang, Xiao-Yan Zhang, Xiao-Yan Zhang, Zhi-Ming Huang, Zhi-Ming Huang, Jie Xiong, Jie Xiong, } "Image classification by multi-instance learning with base sample selection", Proc. SPIE 8335, 2012 International Workshop on Image Processing and Optical Engineering, 833519 (15 November 2011); doi: 10.1117/12.917409; https://doi.org/10.1117/12.917409
PROCEEDINGS
7 PAGES


SHARE
RELATED CONTENT

An approach to image clustering and retrieval
Proceedings of SPIE (April 30 2018)
Character recognition by algebraic invariant matrix
Proceedings of SPIE (October 01 1992)
Discovering salient characteristics of authors of artworks
Proceedings of SPIE (February 16 2010)
Research on methodology of image semantic understanding
Proceedings of SPIE (March 05 2008)

Back to Top