Image Classification Method Based on Weighted Visual Language Model for Saliency Map
Wang Tingji · Jisuanji gongcheng · 2015
At the parameter estimation stage of the image classification method based on the traditional Visual Language Model(VLM),the distribution of visual words is usually analysed via maximum likelihood estimation,which ignores the adverse effect of image background noise. In view of the problem,an image classification method of weighted VLM for saliency map is put forward. The salient regions and background regions are extracted via saliency detection algorithm based on visual attention,the visual documents of images with salient labels are constructed,and the salient weights and conditional probability are estimated in the training phase. The images are classified with weighted VLM for saliency map. Experimental results show that,this method can effectively reduce the influence of image background noise,and enhances the discrimination performance of visual words,so as to improve the performance of image classification based on VLM.