作者: Chao Wang , Jing Zhang , Li Zhuo , Xin Liu
关键词:
摘要: A compressed pornographic image recognition method is proposed by using incremental learning. For describing image, visual words are created from low-resolution (LR) reconstructed the stream of image. Covering algorithm utilized to train and recognize in order build initial classification model At last, learning adopted continuously adjust rules new samples. The experimental results show that for has higher rate as well costs less time.