Development of a face verification system for a home service robot

作者：Yun Woo Han^*; Kim Do Hyung; Yoon Ho Sub; Cho Young Jo

来源：Advanced Robotics, 2008, 22(6-7): 749-760.

DOI：10.1163/156855308X305308

• robot
• face verification
• principle component analysis
• universal background model
• Gaussian mixture model

The robot providing services based on interaction with a human needs a fast verification process that knows whether the user belongs to a family or a guest group for providing the differentiated services to each group member. In this paper, we developed a verification system for one-to-group matching (e.g., user-to-family or user-to-guest) and tested on the database acquired using the robot in an uncontrolled environment. The proposed system consists of three parts: face detection using revised modified census transform and adaboost, feature extraction using multiple principle component analysis, and verification using the aggregating Gaussian mixture model-universal background model. Experimental results show that the proposed system works well on the deteriorated database with a high speed.

• 出版日期2008