Region-based blood color detection and its application to bloody image filtering

Along with the widespread use of the World Wide Web, violent contents have affected our daily life. Although there are some investigations about violence video detection, few methods touch on the problem of violent and gory image detection. In this paper, we propose a region-based blood color detection algorithm. We first extract color and texture features from the detected bloody region of an image. We extract features of the whole image according to the global and local method. These features are fed into the SVM classifier. Experimental results have demonstrated the effectiveness of our proposed algorithm.