Multi Class Adult Image Classification Using Neural Networks

As the Internet became popular, the volume of digital multimedia data is exponentially increased in all aspects of our life. This drastic increment in multimedia data causes unwelcome deliveries of adult image contents to the Internet. Consequently, a large number of children are wide-open to these harmful contents. In this paper, we propose an efficient classification system that can categorize the images into multiple classes such as swimming suit, topless, nude, sexual act, and normal. The experiment shows that this system achieved more than 80% of the success rate. Thus, the proposed system can be used as a framework for web contents rating systems.