Detection of Pornographic Digital Images

In this paper a novel algorithm to detect explicit content or pornographic images is proposed using the transformation from the RGB model color to the YCbCr or HSV color model, moreover using the skin detection the image is segmented, finally the percentage of pixels that was detected as skin tone is calculated. The results obtained using the proposed algorithm are compared with two software solutions, Paraben’s Porn Detection Stick and FTK Explicit Image Detection, which are the most commercial software solutions to detect pornographic images. A set of 800 images, which 400 pornographic images and 400 natural images, is used to test each system. The proposed algorithm carried out identify up to 68.87% of the pornographic images, and 14.25% of false positives, the Paraben’s Porn Detection Stick achieved 71.5% of recognizing but with 33.5% of false positives, and FTK Explicit Image Detection achieved 69.25% of effectiveness for the same set of images but 35.5% of false positives. Finally the proposed algorithm works effectively to carry out the main goal which is to apply this method to forensic analysis or pornographic images detection on storage devices. Keywords—Explicit Content, Pattern Recognition, Skin Detection, The YCbCr and the HSV color models.