Abstract: A novel image categorization algorithm based on improved sparse coding model for image representation and Random Forests for image classification was proposed. Firstly SIFT (ScaleInvariant Feature Transform) descriptors were extracted from images. Then sparse coding was adopted to train a visual dictionary and convert SIFT descriptors into sparse vectors. An efficient pooling method was employed to merge the sparse vectors in each grid of image and a pooled sparse vector was formed to represent the grid. According to the position of grids, the pooled sparse vectors were combined to form a single sparse vector for representing an image. Secondly Random Forests, a multiclass classifier was employed to classify sparse vectors which represent images. The experimental results show that the algorithm outperforms several stateofthearts in image categorization.  

Key words: image categorization, sparse coding, random forests, visual word, BoW model