A Detection Method of Palmprint Principal Lines Based on Local Minimum Gray Value and Line Following

Abstract

A novel detection method of palmprint principal lines is proposed, which employs the priori knowledge of statistical properties about palm lines. Firstly, considering the particular direction of principal lines and the feature of their valley type edges, only one directional template of 45 degrees based on local minimum gray value is used to segment palmprint, and it can significantly reduce the excessive noise that multi-directional templates may bring. Next, a following method of the principal lines is employed, and another linking algorithm for the broken lines is devised. Compared with the existing methods, this scheme can limit principal lines detection in a small area, and trace the principal lines in a certain direction, thus it can effectively avoid the problem of blind searching and enhance the robustness. Finally, an index called "extraction rate"(ER) is defined to evaluate the effect of the proposed approach. The experiments based on the palmprint database of Hong Kong University of Science and Technology (HKUST) show that the extraction rate is 86.67%, and the principal lines are complete. Therefore this method can provide a good support for palmprint identification or other multimodal person identification.