SWU-OFDB: A database for occluded face detection research

Abstract

In the present, surveillance system using camera is popular and widely used in security zone. There are many time that the camera can record thief during committing a crime. While, chance of crime or terrorism occurring is also increased. Thief usually covers their face with an obstacle such as helmet, sunglasses, or mask to blind their face identity. Therefore, to prevent the crime, and improve safety of life, a method to detect person who covering their face in security zone is necessary. However, for testing performance of a method, an occluded face database collected from multi-viewpoints of face is needed. This paper presents overview of current occluded face detection method and the Srinakharinwirot University Occluded Face Database (SWU-OFDB). General process of existing method concerning occluded face detection, and example research in each process for occluded face detection are also explained. © 2014 IEEE.