The new technique is proposed for Re-ranking based image categorization using saliency driven nonlinear diffusion filtering. The image categorize based on Re-ranking, using multi-scale information fusion based on the original image. The foreground features, which are important for image categorization and the background image regions, whether considered as contexts of the foreground or noise to the foreground, can be globally handled by fusing information from different scales. To preserve the foreground feature and deal effectively with background use saliency driven nonlinear diffusion filtering and Re-ranking categorize the image.