Comparative Study of Latent Fingerprint Image Segmentation Techniques Based on Literature Review

Abstract

Latent fingerprints are the fingerprints that are left by the criminals unintentionally on items touched by the fingers. These types of fingerprints are not often directly visible by naked eyes. Segmentation is a very important part of the fingerprint identification system (AFIS). The fingerprint segmentation algorithms separate the foreground (friction ridge pattern) from background. In this paper different segmentation algorithms are presented that are DTV, ADTV, ATV, Ridge Template Correlation method, Segmentation based on statistical characteristics of gray and orientation field information theory, Adaptive Latent Fingerprint Segmentation using Feature Selection and Random Decision Forest Classification, Latent Fingerprint Image Segmentation using Fractal Dimension Features and WELME are discussed and compared their performance. This study evaluates the effectiveness, advantages, limitations and applications of various segmentation methods that are being used in latent fingerprinting segmentation techniques.