Authors: Zhang Guoliang; Wang Zhanni; Liu Yong
Addresses: Computer Science and Technology Institute, Huaqiao University, Xiamen 361021, China ' Computer Science and Technology Institute, Huaqiao University, Xiamen 361021, China ' Computer Science and Technology Institute, Huaqiao University, Xiamen 361021, China
Abstract: A humanoid head robot system with some specific expression was comprehensively modelled in this paper by interactive visual expression analysis between the operator and the robot. Firstly, based on the open source hardware of Fritz humanoid head robot, an integrated emotional expression mapping model was proposed by extending the functional module of voice, vision and so on. Secondly, to achieve a natural interaction between the expression robot and the operator, the integrated visual analysis is mainly utilised for emotion calculation. Haar-like featured cascade classified algorithm is utilised for automatic face location. Gabor feature is filtered and down-sampled as the emotional expression feature. To deal with unreasonable samples distribution in traditional eigenface, an optimised eigenvectors's selection is built by comparing projection reconstructed images with original image. Finally, the classification of with-in class PCA with energy dimension is served as expression recognition. The experimental results on JAFFE database and real-time video image demonstrate the feasibility and effectiveness of the proposed modelling scheme.
Keywords: expression robot; humanoid head robot; Gabor feature; eigenface; feature reconstruction.
International Journal of Internet Manufacturing and Services, 2018 Vol.5 No.2/3, pp.279 - 296
Received: 17 Jun 2017
Accepted: 08 Oct 2017
Published online: 22 May 2018 *