Prototype of Pornographic Image Detection with YCbCr and Color Space (RGB) Methods of Computer Vision

Latest Internet and technology developments have brought in various changes in social interaction, specifically for the teenager to middle age people that excessively using social media. One of the negative effect of this interaction model change is the increase of wide spreading of contents that contain pornographic elements. The pornographic contents can be accessed from social media such as Facebook, Instagram, Path, Whatsapp and Youtube. The purpose of this study is to develop a system for identifying pornographic image content by designing a prototype of pornographic image detection using computer vision's YCbCr and color space (RGB) method by filtering the algorithm of previous researchers, measuring instruments to determine the accuracy of prototypes using the Confusion Matrix algorithm. The experiments using YCbCr produced an accuracy rate of 76%, precision value 82.142%, and recall value 76.666% from 50 random datasets of pornographic positive and non-pornographic images, while the RGB method produced an accuracy rate of 43.42%, precision value 44.23%, and recall value 50.17% from 146 random dataset of pornographic positive, non-pornographic and semi-pornographic images.