DIGITAL IMAGE PROCESSING USING YCBCR COLOUR SPACE AND NEURO-FUZZY TO IDENTIFY PORNOGRAPHY
Cepy Slamet, Yana Aditya Gerhana, Meta Barokatul KR.

UIN Sunan Gunung Djati Bandung

Abstract

Pornography is a very serious problem in Indonesia besides drugs. This can be proven by the increasing number of access to pornographic contents websites on the internet. Some research has been carried out to produce deterrence formulas. However, this research flow has not effectively solved the problem. This study posits a formula to identify pornographic content in digital images as an alternative approach to the initial phase of the system preventing access to negative content. The formulation uses the YCbCr colour space to analyse human skin in image objects representing exposed body parts and classification process by the Neuro-Fuzzy approach. The performance of this formula was tested on 100 digital image images of human objects with random categories (normally closed dressed, minimal, and naked) taken from the internet. The test is resulting in a pretty good level of accuracy with 70% for a weight of the overall test data.

Keywords: pornographic contents identification, YCbCr, Neuro-Fuzzy

Topic: Computer Science