Skin Classification for Adult Image Recognition Based on Combination of Gaussian and Weight-KNN

Nowadays, literature has been explored adult image detection automatic which is a replacement for human action in the boring task of moderating online content. One of the mistake scenes with high skin exposure, such as people swimming and get a tan, can be have many wrong alarms. Some condition factors like illumination, occlusion, and poses are more important to image-recognize which any system has to able to recognize. Reasonable amounts of illumination variation between the gallery and probe images need to be taken into account in image recognition algorithms. In the context of image verification, two items are important; illumination variation and skin classification, and these two factors will most likely result in misclassification. There is a lack of research in combining two factors of imaging condition for illuminating and determining skin in image recognition system. The purpose of this paper is to determine and design the proposed scheme to solve illumination variation and integrate with skin classification in image recognition. The proposed method will be analyzed and evaluated based on its performance in terms of accuracy and effectiveness. In this paper, image processing is divided into two phases; preprocessing and image processing. We have used 8,650 images, which are imported from Compaq and Poesia datasets.