HTML5 Icon

SVM +  Several Iterations  =  Better Results!


The success of an image classification algorithm largely depends on how it incorporates local information in the global decision. Popular approaches such as averagepooling and max-pooling are suboptimal in many situations. In this paper we propose Region Ranking SVM (RRSVM), a novel method for pooling local information from multiple regions. RRSVM exploits the correlation of local regions in an image, and it jointly learns a region evaluation function and a scheme for integrating multiple regions. Experiments on PASCAL VOC 2007, VOC 2012, and ILSVRC2014 datasets show that RRSVM outperforms the methods that use the same feature type and extract features from the same set of local regions. RRSVM achieves similar to or better than the state-of-the-art performance on all datasets.



Author = {Zijun Wei and Minh Hoai},

Booktitle = {Proceedings of IEEE Conference on Computer Vision and Pattern Recognition},

Title = {Region Ranking SVMs for Image Classification},

Year = {2016}}


View RRSVM in Another Perspective

HTML5 Icon

The evolution of top-rank regions and region scores:

From left to right:some image regions and corresponding scores (from the base classifier) at iterations 1, 3, 6, and 9.

(a) for each iteration, the regions at ranks 1, 3, 5 and their scores are shown. For clarity, we list the region scores and region weights underneath each image.

(b,c): the scores of the same set of regions through iterations. The regions in (b) are the ones that are ranked 1, 3, 5 at the first iteration. In (c) are the regions that are ranked 1, 3, 5 after convergence.


This project is partially supported by the National Science Foundation Award IIS-1566248 and IIS-1161876.

We thank Yingtao Tian for providing GPU resources. We also thank Yang Wang and Boyu Wang for helpful discussions and useful feedbacks.