29-06-2012, 04:21 PM
A DCT-based Feature Extraction Algorithm for Palm-print Recognition
[attachment=26159]
INTRODUCTION
Palm-prints of a human being possess some major and
minor line structures along with some ridges and wrinkles.
These markers, as remain stable throughout the lifetime of a
person, are treated as unique biometric features for secure
authentication and identification. Among different
categories of biometric recognition systems, the palm-print
based scheme has become very promising and reliable
because of its robustness against movements of palm, ease
of handling with low resolution images, low memory
requirement, less time consumption and cost-effectiveness.
Most of the palm-print recognition methods primarily
employ three types of feature extraction algorithms, such as
line-based, texture-based, and statistic-based [1]. In the linebased
feature extraction schemes, generally, different edge
detection methods are used to extract palm lines (principal
lines, wrinkles, ridges, etc.) [2]-[4]. The extracted edges,
either directly or being represented in other formats, are
used for matching.
PROPOSED METHOD
DCT Domain Spectral Feature Extraction
In comparison to person recognition based on face or
voice biometrics, palm-print based recognition becomes very
difficult even for a human being. For any type of biometric
recognition, feature extraction is an important task, which
directly dictates the recognition accuracy. As far as palmprint
recognition is concerned, obtaining a significant feature
space with respect to the spatial variation in a palm-print
image is very crucial. In particular, extracting a unique
feature of a palm-print in the spatial domain would be much
difficult as it consists not only some major and minor line
structures, but also some ridges, wrinkles, and singular
points.
EXPERIMENTAL RESULTS
Extensive simulations are carried out in order to demonstrate
the effectiveness of the proposed method for palmprint
recognition using proposed feature vectors. We
investigate the recognition accuracy for the palm-print
images of two well-known databases. The performance of
the proposed method in terms of recognition accuracy is
obtained and compared with those of some recent methods
[9], [10]. Moreover, we investigate the effect of variation of
the widths of the narrow-bands upon the palm-print
recognition accuracy.
CONCLUSIONS
In the proposed DCT-based palm-print recognition
scheme, instead of operating on the entire palm-print image
at a time, dominant spectral features are extracted separately
from each of the narrow-width band obtained by image -
segmentation. It has been found that the proposed feature
extraction scheme offers two-fold advantages. First, it can
precisely capture local variations that exist in the palm-print
images, which plays an important role in discriminating
different persons. Second, it utilizes a very low dimensional
feature space for the recognition task, which ensures lower
computational burden.
[attachment=26159]
INTRODUCTION
Palm-prints of a human being possess some major and
minor line structures along with some ridges and wrinkles.
These markers, as remain stable throughout the lifetime of a
person, are treated as unique biometric features for secure
authentication and identification. Among different
categories of biometric recognition systems, the palm-print
based scheme has become very promising and reliable
because of its robustness against movements of palm, ease
of handling with low resolution images, low memory
requirement, less time consumption and cost-effectiveness.
Most of the palm-print recognition methods primarily
employ three types of feature extraction algorithms, such as
line-based, texture-based, and statistic-based [1]. In the linebased
feature extraction schemes, generally, different edge
detection methods are used to extract palm lines (principal
lines, wrinkles, ridges, etc.) [2]-[4]. The extracted edges,
either directly or being represented in other formats, are
used for matching.
PROPOSED METHOD
DCT Domain Spectral Feature Extraction
In comparison to person recognition based on face or
voice biometrics, palm-print based recognition becomes very
difficult even for a human being. For any type of biometric
recognition, feature extraction is an important task, which
directly dictates the recognition accuracy. As far as palmprint
recognition is concerned, obtaining a significant feature
space with respect to the spatial variation in a palm-print
image is very crucial. In particular, extracting a unique
feature of a palm-print in the spatial domain would be much
difficult as it consists not only some major and minor line
structures, but also some ridges, wrinkles, and singular
points.
EXPERIMENTAL RESULTS
Extensive simulations are carried out in order to demonstrate
the effectiveness of the proposed method for palmprint
recognition using proposed feature vectors. We
investigate the recognition accuracy for the palm-print
images of two well-known databases. The performance of
the proposed method in terms of recognition accuracy is
obtained and compared with those of some recent methods
[9], [10]. Moreover, we investigate the effect of variation of
the widths of the narrow-bands upon the palm-print
recognition accuracy.
CONCLUSIONS
In the proposed DCT-based palm-print recognition
scheme, instead of operating on the entire palm-print image
at a time, dominant spectral features are extracted separately
from each of the narrow-width band obtained by image -
segmentation. It has been found that the proposed feature
extraction scheme offers two-fold advantages. First, it can
precisely capture local variations that exist in the palm-print
images, which plays an important role in discriminating
different persons. Second, it utilizes a very low dimensional
feature space for the recognition task, which ensures lower
computational burden.