Informative Patches Sampling for Image Classification by Utilizing Bottom-up and Top-down Information

概要

論文の詳細を見る
In image classification based on bag of visual words framework, image patches used for creating image representations affect the classification performance significantly. However, currently, image patches are sampled mainly based on processing low-level image information or just extracted regularly or randomly. These methods are not effective. In this report, we propose to explore both bottom-up information through processing low-level image information and top-down information through utilizing the statistical properties of training image grids to extract image patches. In the proposed work, an input image is divided into regular grids, each of which is evaluated based on its bottom-up information and/or top-down information. Subsequently, a saliency value is assigned to every grid based on the evaluation results, so that a saliency map can be created for the input image. Finally, the sampling of image patches is performed on the basis of the obtained saliency map. Furthermore, we propose a method to fuse the two kinds of information. The proposed methods are evaluated on both object categories and scene categories. Experiment results demonstrate their effectiveness.
一般社団法人電子情報通信学会の論文
2012-02-02