Evaluating bag-of-visual-words representations scene

Based on keypoints extracted as salient image patches, an image can be described as a “bag of visual words ” and this representation has been frequently used in the classification of imagery data. The representation choices regarding the dimension, selection, and weighting of visual words are cru-cial to the classification performance but have not been thor-oughly studied in existing works. Given the analogy between this image representation and the bag-of-words representa-tion of text documents, we apply techniques widely used in text categorization, including term weighting, stop word re-moval, feature selection, to generate image representations that differ in the dimension, selection, and weighting of vi-sual words. The impact of these representations choices to scene classification is studied through extensive experiment