Aditya Nigam

An efficient finger-knuckle-print based recognition system fusing SIFT and SURF matching scores

This paper presents a novel combination of local-local information for an efficient finger-knuckle-print (FKP) based recognition system which is robust to scale and rotation. The non-uniform brightness of the FKP due to relatively curvature surface is corrected and texture is enhanced. The local features of the enhanced FKP are extracted using the scale invariant feature transform (SIFT) and the speeded up robust features (SURF). Corresponding features of the enrolled and the query FKPs are matched using nearest-neighbour-ratio method and then the derived SIFT and SURF matching scores are fused using weighted sum rule. The proposed system is evaluated using PolyU FKP database of 7920 images for both identification mode and verification mode. It is observed that the system performs with CRR of 100% and EER of 0.215%. Further, it is evaluated against various scales and rotations of the query image and is found to be robust for query images downscaled upto 60% and for any orientation of query image.

Iris Segmentation Using Improved Hough Transform

This paper presents an efficient iris segmentation algorithm. This paper uses an improved circular Hough transform to detect inner boundary and the circular integro-differential operator to detect the outer boundary of iris from a given eye image. Search space of the standard circular Hough transform is reduced from three dimensions to only one dimension, which is the radius. Local gradient information is used to improve time and efficiency of Hough transform. This algorithm has been tested on the publicly available CASIA 3.0 Interval database consisting of 2639 images of 249 subjects and CASIA 4.0 Lamp database consisting of 16,212 images of 411 subjects. It also provides error categorization for wrong segmentation, as well as a study on parametric influences on error. Parameterized error analysis helps to set parameters intelligently boosting up the segmentation accuracy as high as 99.8% on the Interval database and 99.7% on the Lamp database.

Finger knuckleprint based recognition system using feature tracking

This paper makes use of finger knuckleprints to propose an efficient biometrics system. Edge based local binary pattern (ELBP) is used to enhance the knuckleprint images. Highly distinctive texture patterns from the enhanced knuckleprint images are extracted for better classification. It has proposed a distance measure between two knuckleprint images. This system has been tested on the largest publicly available Hong Kong Polytechnic University (PolyU) finger knuckleprint database consisting 7920 knuckleprint images of 165 distinct subjects. It has achieved CRR of more than 99.1% for the top best match, in case of identification and ERR of 3.6%, in case of verification.

Iris recognition using consistent corner optical flow

This paper proposes an efficient iris based authentication system. Iris segmentation is done using an improved circular hough transform and robust integro-differential operator to detect inner and outer iris boundary respectively. The segmented iris is normalized to polar coordinates and preprocessed using LGBP (Local Gradient Binary Pattern). The corners features are extracted and matched using dissimilarity measure CIOF (Corners having Inconsistent Optical Flow). The proposed approach has been tested on publicly available CASIA 4.0 Interval and Lamp databases consisting of 2,639 and 16,212 images respectively. It has been observed that the segmentation accuracy of more than 99.6% can be achieved on both databases. This paper also provides error classification for wrong segmentation and also determines influential parameters for errors. The proposed system has performed with CRR of 99.75% and 99.87% with an EER of 0.108% and 1.29% on Interval and Lamp databases respectively.

Comparing human faces using edge weighted dissimilarity measure

This paper proposes a dissimilarity measure that can be used as a distance between two images. It has shown better discriminative power to recognize faces than similar existing variants for discriminating facial images. It gives more weight to the pixels which are often a part of the edge and counts pixels that are unmatched between query and database images. This measure has been tested on a publicly available database as ORL, YALE, CALTEC, BERN and also on a database developed at IIT Kanpur. Experimental results show that the proposed measure achieves a high recognition rate of 99.75% 93.75% 99.03% 98.93% 99.73% for the first likely matched faces on databases ORL, YALE, BERN, CALTECH, IITK respectively. The proposed measure can provide not only effective result against pose and expression variations but also against slight illumination variation.

Four slap fingerprint segmentation

This paper proposes an efficient segmentation algorithm to extract multiple fingerprints from a four slap fingerprint image. The proposed algorithm divides each slap image into non-overlapping square blocks and identifies foreground and background blocks using average pixel intensity in that block. Blocks are joined to find the foreground blocks connected components. Finally clusters containing finger-tips are selected using geometric characteristics of a finger. It has been tested on IITK four slap fingerprint database of 13380 four slap fingerprint images collected from 1115 subjects in 2 sessions. The algorithm has achieved segmentation accuracy of 98.8%.

Fusion of 4-slap fingerprint images with their qualities for human recognition

This paper presents an efficient multimodel bio-metric system based on 4 slap fingerprint images. The system utilizes 4 slap fingerprint scanner to simultaneously collect fingerprints of multiple fingers on a hand in one image. The acquired multi-finger images are first segmented to get individual fingers. Quality of each individual finger is estimated and its minutiae points are extracted. The minutiae points of each individual finger extracted from gallery 4 slap fingerprint image is compared with the corresponding individual finger of the query 4 slap fingerprint image to get matching score of that finger. Matching score between two 4 slap fingerprint images is obtained by fusing matching scores of various fingers along with their respective image quality and relative accuracies. Decision of matching has been taken based on the fused matching score. The system has been tested on two 4 slap fingerprint databases viz IITK-student and IITK-rural containing 1007 and 991 subjects respectively. Both databases are acquired in 2 sessions. The correct recognition rate obtained is 91.00% for IITK-rural database and 99.64% for IITK-student database. Respective EER values are 5.64% and 0.94%.

A New Distance Measure for Face Recognition System

This paper proposes a new powerful distance measure called Normalized Unmatched Points (NUP). This measure can be used in a face recognition system to discriminate facial images. It works by counting the number of unmatched pixels between query and database images. A face recognition system has been proposed which makes use of this proposed distance measure for taking the decision on matching. This system has been tested on four publicly available databases, viz. ORL, YALE, BERN and CALTECH databases. Experimental results show that the proposed measure achieves recognition rates more than 98.66% for the first five likely matched faces. It is observed that the NUP distance measure performs better than other existing similar variants on these databases.

Quality assessment of knuckleprint biometric images

Image quality can play key role in the system performance. The recently introduced knuckleprint biometric has shown promising results, but its quality assessment is difficult because it lacks well defined and structured features as in the case of face or fingerprint. To our knowledge this is the first attempt to automatically assess the quality of knuckleprint images. The quality of knuckleprint images mainly depends upon the vertical line like features, focus, contrast and reflections produced by the camera flash. In this paper an effort has been made to identify, estimate and quantify some of these quality attributes and fuse them to obtain an overall quality score for any knuckleprint image. The largest publicly available PolyU knuckleprint database is used for testing, containing 7920 images. Extensive study is being carried out in order to establish a relationship between image quality and matching performance in order to demonstrate the proposed frameworks utility.

Multiple texture information fusion for finger-knuckle-print authentication system

Abstract This paper proposes a finger-knuckle-print based authentication system by fusing multiple texture features. It contains new algorithms for extracting region of interest (ROI) with the help of curvature Gabor filters, image quality parameters, ROI enhancement using gradient based ordinal relationships, and dissimilarity measure for matching. The proposed system has been tested on the largest publicly available finger-knuckle-print PolyU database consisting of 7920 finger-knuckle-print images obtained from 165 subjects in two sessions. It has shown good performance.