Neural Network-Based Nipple Detection for Obscene Pictures

Self-Organizing Map (SOM) is an unsupervised neural network using for object detection and recognition in numerous image processing applications. This paper presents a methodology to achieve an automated detection of nipples by obscene pictures. The proposed system is composed of a human skin detection using for non-skin removal process based on image processing and a nipple detection using for nipple existence identification based on Self-Organizing Map. The goal is to identify whether the given picture is an obscene picture by detecting the nipple existence in the picture. The proposed system is shown to be effective for a wide range of shades and colors of skin and human configurations. It is validated for detecting a nipple existence for obscene pictures. In terms of scalability, the nipple detection model can be modified to support distributed processing.