作者: Yizhi Liu , Hongtao Xie
DOI: 10.1109/ICCIT.2009.5407272
关键词:
摘要: Pornographic images detection is necessary for us to filter out objectionable information on the Internet. Bag-of-visual-words (BoVW) based pornographic promising because it can compensate defect of traditional approach. However, there are many choices construct visual-words which crucial tradeoff between speed and performance. We propose a novel method constructing SURF (speeded up robust features) in skin regions combining with color moments. The results show that performance better than SIFT (scale-invariant feature transform) our more effective detect existing methods.