Implementation of K-NN based on histogram at image recognition for pornography detection

The development of information technology today has a positive and negative impact. One of the negative impacts is the spread of images containing inappropriate content (porn) uncontrollably so that it can be accessed by users from all walks of life, especially minors. There are several techniques to control the negative impact of the development of information technology, and one of them is by utilizing digital image processing in recognizing and detecting an image containing pornographic content. The technique used in this research is color segmentation on the image with YCbCr color model and classified and look for the similarity of training data with data testing using K-NN algorithm. The results obtained in this study using European and Asian skin color samples show that the prototype model designed can be used to detect pornographic, semi-pornographic or non-pornographic images with 90% accuracy.