ROI Extraction of Palmprint Images Using Modified Harris Corner Point Detection Algorithm

摘要

A new extraction method of palmprint images' region of interesting (ROI) by using a modified Harris corner detection algorithm is discussed in this paper. This method can independently select the corner detection region by using mouse and doesn't need to detect the whole palmprint image. So, the calculation complexity is reduced hardly. At the same time, this method can ensure the uniqueness of detected corners and improve the location accuracy of palmprints. In test, the database of CASIA is used. Compared with the traditional Harris corner detection algorithm, experimental results show that our method can more efficiently and quickly locate palmprint images' ROI, and can efficiently enhance the edge and detail of palmprint images simultaneity.