Domain Adaptation for Holistic Skin Detection

Human skin detection in images is a widely studied topic of Computer Vision for which it is commonly accepted that analysis of pixel color or local patches may suffice. However, we found that the lack of contextual information may hinder the performance of local approaches. In this paper, we present a comprehensive evaluation of holistic and local Convolutional Neural Network (CNN) approaches on in-domain and cross-domain experiments and compare them with state-of-the-art pixel-based approaches. We also propose combining inductive transfer learning and unsupervised domain adaptation methods evaluated on different domains under several amounts of labelled data availability. We show a clear superiority of CNN over pixel-based approaches even without labeled training samples on the target domain and provide experimental support for the superiority of holistic over local approaches for human skin detection.