MUA3D: Malaysian ethnicity recognition

Abstract

From a human face, one can tell that person’s emotion, gender and ethnicity. The ethnicity recognition score contributes to increase the matching score in the face recognition. This paper presents Malaysian ethnicity recognition using a new face database, MUA, developed by a biometric research group, Sarawak Biometric (SARAB). This database contains data from three main ethnic groups in Malaysia mainly Malay, Chinese and Indian. In addition, MUA also contains data for the biggest ethnics in Sarawak, the Ibans or called Sea Dayaks. We used 200 face data from MUA. Several geometric features were extracted from the images. The classifier used is Support Vector Machine whereas the accuracy obtained for facial ethnicity classification is 92.01%.