Palmprint Recognition Based on Complete Direction Representation

Abstract

Direction information serves as one of the most important features for palmprint recognition. In the past decade, many effective direction representation (DR)-based methods have been proposed and achieved promising recognition performance. However, due to an incomplete understanding for DR, these methods only extract DR in one direction level and one scale. Hence, they did not fully utilize all potentials of DR. In addition, most researchers only focused on the DR extraction in spatial coding domain, and rarely considered the methods in frequency domain. In this paper, we propose a general framework for DR-based method named Complete Direction Representation (CDR), which reveals DR by a comprehensive and complete way. Different from traditional methods, CDR emphasizes the use of direction information with strategies of multi-scale, multi-direction level, multi-region, as well as feature selection or learning. This way, CDR subsumes previous methods as special cases. Moreover, thanks to its new insight, CDR can guide the design of new DR-based methods toward better performance. Motived this way, we propose a novel palmprint recognition algorithm in frequency domain. Firstly, we extract CDR using multi-scale modified finite radon transformation (MFRAT). Then, an effective correlation filter, namely Band-Limited Phase-Only Correlation (BLPOC), is explored for pattern matching. To remove feature redundancy, the Sequential Forward Selection (SFS) method is used to select a small number of CDR images. Finally, the matching scores obtained from different selected features are integrated using score-level-fusion.

Existing System

Discriminant DCT Feature Extraction, Support Vector Machine (SVM).

Proposed System

We propose a unified framework for DR-based methods, named complete direction representation (CDR). CDR simultaneously considers comprehensive factors to capture information in multi-scale, multi-direction level, and local regions with an effective feature selection mechanism. While subsuming previous DR-based methods as special cases, CDR provides more insights for understanding them and consequently guides us toward improvement. We propose a novel method in frequency domain based on the CDR-framework. The first step in our method is to extract rich multi-factor information through the multi-scale modified finite radon transformation (MFRAT). Then, an effective correlation filter, Band-Limited Phase-Only Correlation (BLPOC), is used for pattern matching. After that, to remove redundant features, Sequential Forward Selection (SFS) is adopted to select a small number of CDR images to achieve the best recognition performance. Finally, the selected features are integrated using score-level fusion. To the best of our knowledge, the proposed method is the first work that systematically investigate DR-based method in frequency domain.

Architecture

BLOCK DIAGRAM