Segmentation of Lip Print Images Using Clustering and Thresholding Techniques

Abstract

Segmentation process forms a vital component of image processing. The major objective of lip print image segmentation is to separate the image pixels into foreground pixels that contain the region of interest and background pixels that mainly consists of noise. This partition of the original lip print image into various meaningful representations makes it easy to analyze the image. There are various segmentation techniques available that can be used for certain problem statements. But in some cases, many of the existing techniques need to be combined along with our knowledge on the domain to productively solve a problem on image segmentation. In this paper, algorithms to effectively segment the original lip print image into upper and the lower lip are presented. Thresholding and clustering techniques are used to segment the lip print images. Results show that the presented techniques provide good performance and better segmentation results. It also shows that the noise pixels are effectively categorized as the background pixels while the portion consisting of the region of interest is categorized as the foreground pixels effectively.