Abstract:

A large number of vision applications rely on matching keypoints across images. To overcome the shortcomings of the Features from Accelerated Segment Test (FAST ) and the Binary Robust Independent Elementary Features (BRIEF ), we propose an improved image matching algorithm. Firstly, simple mask is applied in the FAST algorithm to extract image keypoints, and scale invariance is achieved by the image pyramid. The sampling pattern in the BRIEF is modified according to the principles of the human visual system, and the keypoints with rotation invariance are achieved by the orientation estimation. Furthermore, the keypoints descriptor similarities are evaluated by using the Hamming distance, which is very efficient to compute. Experimental results show that the proposed algorithm is not only faster to compute and also more robust than other algorithms.

Key words: image matching;FAST;BRIEF;scale invariance;rotation invariance