Skip to Main Content
The technology of face recognition is a well-known subject. This paper presents a face feature extraction method based on MB-LBP. In this method, by means of pre-locating the feature points, a Multi-Block Local Binary Pattern (MB-LBP) is intensively studied and related algorithm implemented. The proposed method is then tested by using Yale face database, and how MB-LBP scale (or size of training set) influences on the face recognition results is explored. Evaluation results confirm that the implemented algorithm is both valid and effective under certain conditions.
Power Engineering and Automation Conference (PEAM), 2011 IEEE (Volume:3 )
Date of Conference: 8-9 Sept. 2011