Sid Ray

Fingerprint Recognition System Using Hybrid Matching Techniques

With an increasing emphasis on the emerging automatic person identification application, biometrics based, especially fingerprint-based identification, is receiving a lot of attention. This research developed an automatic fingerprint recognition system (AFRS) based on a hybrid between minutiae and correlation based techniques to represent and to match fingerprint; it improved each technique individually. It was noticed that, in the hybrid approach, as a result of an improvement of minutiae extraction algorithm in post-process phase that combines the two algorithms, the performance of the minutia algorithm improved. An improvement in the ridge algorithm that used centre point in fingerprint instead of reference point was also observed. Experiments indicate that the hybrid technique performs much better than each algorithm individually.

A Comparison of Relevance Feedback Strategies in CBIR

Relevance feedback (RF) is considered to be very useful in reducing semantic gap and thus enhancing accuracy of a Content-Based Image Retrieval system. In this paper, we have given a brief overview of research done in this area with an emphasis on feature re-weighting approach, a popular RF technique. We have also discussed an instance-based approach that has been introduced very recently. We considered image retrieval as a dichotomous classification problem and compared performances of the two RF strategies with four different datasets, with number of images ranging from 1000 to 19511.

Optimization of Iris Codes for Improved Recognition

The texture of the iris is commonly represented as an iris code in iris recognition systems. While several approaches have been presented for generating iris codes, relatively few comparison techniques have been proposed. In this paper, we take advantage of the availability of several frames from an iris video to create a single optimized iris code. This is achieved by performing both row-wise and column-wise optimization of iris codes. Inconsistent bits are accurately detected and masked in the final iris code. Our experiments demonstrate that by exploiting variations within the comparison scores of different rows and columns of N frames, we are able to derive the number of consistent bits in the final iris code thereby resulting in significant improvement in recognition performance. We compare our algorithm with well-known methods, namely, Fragile bit masking, Signal fusion and, two Score Fusion techniques. Experimental results on a dataset of 986 iris videos show that the proposed method is encouraging and comparable to the best algorithms in the current literature. To our knowledge, this is the first work that makes use of the best rows and columns from different frames in an iris video to improve performance.

Model-based pupil and iris localization

The iris is the most accurate biometric to date and its localization is a vital step in any iris recognition system. Iris localization can be considered as the search for the demarcation points, or step change in intensity at its boundaries. A failed localization will lead to incorrect iris segmentation and eventually to poor recognition. In the first stage, we proceed with the elimination of reflection and the reduction of lighting variations in eye images. In the second stage of our proposed system, radii and locations of the pupil and iris are obtained by maximizing the convolution of the image with a toroidal 2-D filtering shape derived from the Petrou-Kittler 1-D filter. Such a novel approach delivers robust localization of the inner and outer iris boundaries. We tested our system on a large dataset of poor quality eye images with substantial occlusions, illumination and defocus and the proposed algorithm is found to be robust and accurate.

Robust Video Based Iris Segmentation System in Less Constrained Environments

One of the key challenges in traditional iris recognition systems is that they require substantial user cooperation. Several restrictions are imposed on positioning and motion of the subject during the image acquisition process so that an image of high quality can be captured. On the other hand, videos captured at a distance and on the move are less intrusive and more appealing to users. However, this extra convenience comes at a cost. Such videos suffer from significant degradation and are often of poor quality compared to images captured in controlled environments. In this work, we present a video based iris segmentation system for processing of images taken in less constrained environments. In the first part, frame alignment of face videos is performed for reliable and efficient extraction of the eye regions in Near Infrared (NIR) videos. In the second section, we propose a new iris segmentation method aimed particularly at eye images captured in challenging environments. Reflections and out of frame iris regions are in-painted. A region based segmentation method is proposed for accurate eyelid detection in images with variable illumination and significant blur. Eyelashes are divided into two categories and eliminated. Experiments carried out on the Multiple Biometric Grand Challenge (MBGC) dataset demonstrate that the proposed system achieves higher accuracy than other recent state of the art video based iris segmentation techniques developed for less constrained environments.

Automated Selection of Optimal Frames in NIR Iris Videos

A relatively new trend in the iris biometric area is the use of videos as a capturing device. Frame by frame approach is richer in information and gives more flexibility as opposed to the use of traditional still images. However, the quality, shape and size of the iris may vary from one frame to another. In this paper, we propose a new technique for selecting the best frames in an iris video. Taking advantage of the temporal correspondence in iris frames, we classify iris videos into 3 categories, namely Adequate, Motion Constrained and Time Constrained. Frames with blinks and off-angle gaze are eliminated using frame averaging and correlation. Quality factors, namely motion blur, out of focus, translational motion and lighting present in iris videos are detected and their effect on recognition performance is investigated. Experimental results are carried out on both the MBGC NIR Iris Video and the MBGC NIR Iris Still datasets from the National Institute for Standards and Technology (NIST). Firstly, this work demonstrates that the proposed optimal frame selection technique in NIR Iris Videos leads to significant improvement in recognition performance. Secondly, the performance of NIR Iris Still images vs. NIR Iris Videos is compared. Thirdly, we show that interoperability between iris frames and iris images in an iris recognition system affects performance. Finally, the computational time and the elimination of noisy frames at each stage using the proposed method are examined.

Computer Aided Analysis of Dental Radiographic Images

This paper is a result of a fruitful cooperation between the computer science and the dental diagnosis experiences. The study presents a new approach of applying computer algorithms to radiographic images of dental implantation used for bone regeneration. We focus here only on the contribution of the computer assistance to the clinical research as the periodontal therapy is beyond the scope of this paper. The proposed system is based on a pattern recognition approach, directed to recognize density changes in the intra-bony affected areas of patients. It comprises different modules with new algorithms specially designed to treat the patients’ radiographic images more accurately. The system includes digitizing, detecting the complicated region of interest (ROI), defining reference area to correct any projection discrepancy of the follow up images, and finally to extract the distinguishing features of the ROI as a basis for determining the rate of new bone density accumulation. This study is applied to two typical dental cases for a patient who received two different operations. The results are very encouraging and more accurate than traditional techniques reported before.

Effect of Finite Sample Size in Content-Based Image Retrieval

Finite sample size has always been a problem in determining the retrieval accuracy of a Content-Based Image Retrieval (CBIR) system. Though a good amount of research has been done in the statistical pattern recognition field, no such effort is shown in relation to CBIR. In this paper, we considered image retrieval as a dichotomous classification problem and studied the effect of sample size on the retrieval accuracy. We reported experimental results and analysis with two different image databases of size 2000 and 500, both having 10 semantic categories. For both data sets, we showed the variation of precision with sample size. We also studied the effect of sample size on retrieval accuracy as Relevance Feedback (RF) is applied. For both data sets, the nett improvement in precision with RF increases with sample size.

Improved Approximate Multiple-Pattern String Matching using Consecutive N-Grams

matching is to find all the occurrences of a given pattern in a large text, the strings being sequence of characters drawn from finite alphabet set. Multiple-Pattern string matching problem involves detection of all the patterns of the Multiple-Pattern set in the text. Shift OR algorithm which we call as the Standard Shift OR algorithm uses the concept of Bit Parallelism to perform approximate string matching. The algorithm as the name suggests performs approximate string matching which means that it finds out some false matches besides detecting correct matches. In other words the algorithm behaves as a filter. In this paper a modification of the standard Shift OR is proposed to improve the filtering efficiency of the standard Shift OR algorithm using the consecutive N-Grams of the patterns of the multiple-pattern set. The proposed method reads N characters of the text at once as compared to a single character in the standard Shift OR algorithm. The number of false matches reduces besides increasing the speed of matching. Extensive experiments have been performed with the algorithm on text and pattern of variable size and the results are compared with the standard Shift OR algorithm.

Isolation of lung cancer from inflammation

In this paper we propose an efficient new algorithm for making an intelligent system for isolating the lung cancer from the inflamed region using needle biopsies. The best way among the cancer treatments, surgery, is the way that can be used for the removal of a malignant tumor in an operation. It is most effective when a cancer is small and localized. Identification and removal of the cancer cells in their earliest formation are very much important. Almost all of the diagnostic laboratories in the world use experts to identify the suspected cells of the lung tumors under microscope. Due to the smaller number of experts, the proposed method, derived based on image contour analysis, has an important significance to replace the manual methods by an intelligent system.