Filtering of NSFW Images in the Information Flow

Convolutional Neural Networks (CNN) shows state of the art results on variety of tasks. The paper presents the scheme how highly accurate CNN (97.8% on the test set) for filtering not suitable or safe for work (NSFW) images was created. The present research focuses on investigating questions concerning identifying NSFW pictures with nudity by neural networks. One of the main feature of the present work is considering of the NSFW class of images not only in terms of natural human nudity but also include cartoons and other drawn pictures containing obscene images of the primary sexual characteristics. Another important considered issue is creating procedure and collecting representative dataset for the problem by it. The research includes the review of existing NSFW detection methods which are provided by traditional machine learning techniques and quite new neural networks based approaches. In addition several important problems in NSFW pictures filtering are considered in the study. The results of the work are dataset and its collecting procedure, trained neural network and comparison with other existing approaches.