Palm lines extraction using PCNN and image data field
Yanxia Wang, Jianmin Zhao, Guang‐Hua Sun, Hui Wang, Xin Chen, Dewu Xu · 2013
In this paper, an approach of using image data field and pulse-coupled neuron network (PCNN) to fast extract palm lines is proposed for online palmprint images. Each pixel in an enhanced palmprint image is seen as a particle with the mass, which produces a data field. The data field is introduced to map the enhanced palmprint image from grayscale space to the corresponding potential space. By selecting the relative mass, a relative image data field is obtained. Then the enhanced palmprint image and its relative data field are input into two PCNNs with different parameters, separately. Their outputs are two binary images, and each has complementary advantages and disadvantages. In order to extract palm lines, a non-connectedness value (NCV) is defined and used to fusion the two binary images. At last, the morphological operators are used to remove noise and isolated points. The experimental results indicate that the speed of palm-line extraction promotes greatly and it can satisfy the practical requirements.