電子情報通信学会総合大会講演要旨
D-12-69
Robust Recognition of Multiple Particular Users’ Group for Autonomous Guiding Robot
○Bin Zhang・Tomoaki Nakamura・Masahide Kaneko(The Univ. of Electro-Communications)
In this paper, we proposed a new robust recognition method of multiple particular users for an autonomous group guiding robot. By integrating multiple recognition results of height, color, gait and face information, we get a better recognition rate than any individual information. The height and color histogram based recognition mainly work on deleting the candidates that have obvious differences. The gait information gotten from laser sensor and Local Binary Pattern (LBP) histogram matching based face recognizer mainly work on calculating the similarity, never for rejecting the candidates. The best recognition result is accepted with the joint probability.