Real-Time Object Segmentation Architecture for Household Robot Applications

다국어 초록 (Multilingual Abstract)

This paper presents hardware architectures of the real-time object segmentation of the noisy disparity map which can be obtained by using stereo matching algorithm. The disparity map contains lots of noise due to various causes, and it has to be refined by some noise filtering methods to make it useful for the object segmentation. In our approach, refinement method based on noise elimination technique is adopted for improvement of the disparity map quality. And the projection-based region merging method is used for object segmentation. The proposed algorithms are implemented in real time FPGA board. Results of real-time test show that our approach works precisely and its performance do not violate realtime qualifications. The developed real-time object segmentation system could be applied for face recognition, object tracking, and other applications with the support of proper embedded software.