Dosung Ahn

Automatic alignment of fingerprint features for fuzzy fingerprint vault

Biometrics-based user authentication has several advantages over traditional password-based systems for standalone authentication applications. This is also true for new authentication architectures known as crypto-biometric systems, where cryptography and biometrics are merged to achieve high security and user convenience at the same time. Recently, a cryptographic construct, called fuzzy vault, has been proposed for crypto-biometric systems. This construct aims to secure critical data(e.g., secret encryption key) with the fingerprint data in a way that only the authorized user can access the secret by providing the valid fingerprint, and some implementations results for fingerprint have been reported. However, all the previous results assumed that fingerprint features were pre-aligned, and automatic alignment in the fuzzy vault domain is a challenging issue. In this paper, we perform the automatic alignment of fingerprint features by using the geometric hashing technique which has been used for model-based object recognition applications. Based on the preliminary experimental results, we confirm that the proposed approach can align fingerprint features automatically in the domain of the fuzzy vault and can be integrated with any fuzzy fingerprint vault systems.

Performance evaluation of watermarking techniques for secure multimodal biometric systems

In this paper, we describe various watermarking techniques for secure user verification in the remote, multimodal biometric systems employing both fingerprint and face information, and compare their effects on user verification and watermark detection accuracies quantitatively. To evaluate the performance of watermarking for multimodal biometric systems, we first consider possible two scenarios – embedding facial features into a fingerprint image and embedding fingerprint features into a facial image. Additionally, to evaluate the performance of dual watermarking for secure biometric systems, we consider another two scenarios – with/without considering the characteristics of the fingerprint in embedding the dual watermark. Based on the experimental results, we confirm that embedding fingerprint features into a facial image can provide superior performance in terms of the user verification accuracy. Also, the dual watermarking with considering the characteristics of the fingerprint in embedding the dual watermark can provide superior performance in terms of the watermark detection accuracy.

An ultra-low memory fingerprint matching algorithm and its implementation on a 32-bit smart card

We propose an ultra-low memory fingerprint matching algorithm and implement it on a 32-bit smart card. We first evaluated both the number of instructions executed and memory requirement of each step of a typical fingerprint matching algorithm. Then, we developed a memory-efficient algorithm for the most memory-consuming step (alignment) by doing more computations under the restriction of the real-time requirement. Our experimental results show that the proposed algorithm can reduce the required memory space by a factor of 62 and can be executed in real-time on a 32-bit smart card.

Matching with Secure Fingerprint Templates Using Non-invertible Transform

This paper presents an approach to generate secure minutiae information in fingerprint templates. Using triplets of minutia points, useful geometrical properties can be derived from minutia triplets to hide their coordinates, direction of minutia, and scale. This non-invertible transform enables matching from the results in the anonymous domain to fight against malicious attacks to biometrics. The goal is to achieve non-invertibility of the transformed templates without loss of discriminating capability. Proposed features can match two templates without alignment under low time complexity and memory requirement. The proposed approach is proven effective using the FVC2002 database to compare matching performances. The best result shows 3.61% of equal-error rate for the database DB2_A.

Access control system with high level security using fingerprints

Biometric based applications guarantee for resolving numerous security hazards. As a method of preserving of privacy and the security of sensitive information, biometrics has been studied and used for the past few decades. Fingerprint is one of the most widely used biometrics. A number of fingerprint verification approaches have been proposed until now. However, fingerprint images acquired using current fingerprint input devices that have small field of view are from just very limited areas of whole fingertips. Therefore, essential information required to distinguish fingerprints could be missed, or extracted falsely. The limited and somewhat distorted information are detected from them, which might reduce the accuracy of fingerprint verification systems. In the systems that verify the identity of two fingerprints using fingerprint features, it is critical to extract the correct feature information. In order to deal with these problems, compensation of imperfect information can be performed using multiple impressions of enrollees fingerprints. In this paper, additional three fingerprint images are used in enrollment phase of fingerprint verification system. Our experiments using FVC 2002 databases show that the enrollment using multiple impressions improves the performance of the whole fingerprint verification system.

Fusion for multimodal biometric identification

In this paper, we investigate fusion methods for multimodal identification using several unimodal identification results. One fingerprint identification system and two face identification systems are used as fusion sources. We discuss rank level and score level fusion methods. Whereas the latter combines similarity scores, the other one combines the orders of the magnitudes of the similarity scores. For rank level methods, Borda Count and Bayes Fuse are considered and, for score level methods, Sum Rule and Binary Classification Approach are considered. Especially, we take a more detailed look at Binary Classification Approach, which simplifies a multiple class problem into a binary class problem. Finally, we compare experimental results using the fusion methods in different combinations of the sources.

An Experimental Study on Measuring Image Quality of Infant Fingerprints

The performance of fingerprint recognition algorithms de- pends on quality of input fingerprint images. This dependency becomes more apparent for infant fingerprints that are smaller than adult finger- prints. This paper studies various measurements of fingerprint images, and compares the image qualities between infant and adult fingerprints. Three quality measures for fingerprint images are implemented, and applied to infant and adult fingerprint images. The results can be util- ized for ensuring better performance of fingerprint recognition systems against infant fingerprints.

Fingerprint-based authentication for USB token systems

In the modern electronic world, the authentication of a person is an important task in many areas of day-to-day. Using biometrics to authenticate a person’s identity has several advantages over the present practices of Personal Identification Numbers (PINs) and passwords. To gain maximum security in the authentication system using biometrics, the computation of the authentication as well as the store of the biometric pattern has to take place in the security token(e.g., smart card, USB token). However, there is an open issue of integrating biometrics into the security token because of its limited resources(processing power and memory space). In this paper, we describe our implementation of the USB token system having 206MHz StrongARM CPU, 16MBytes Flash memory, and 1MBytes RAM. Then, we describe a fingerprint enrollment algorithm that can check false minutiae detected and true minutiae missed by using multiple impressions. Also, to meet the memory space specification and processing power of the security token in fingerprint verification algorithm, we describe a memory-efficient alignment algorithm. Based on experimental results, we confirmed that the RAM requirement of the proposed algorithm is about 16 KBytes, and the Equal Error Rate(EER) is 1.7%. Therefore, our fingerprint authentication algorithm can be executed in real-time on the developed USB token.

User Enrollment Using Multiple Snapshots of Fingerprint

As a method of preserving of privacy and the security of sensitive information, biometrics has been studied and used for the past few decades. A number of fingerprint verification approaches have been proposed until now. However, fingerprint images acquired using current fingerprint input devices that have small field of view are from just very limited areas of whole fingertips. Therefore, essential information required to distinguish fingerprints could be missed or extracted falsely. The limited and somewhat distorted information are detected from them, which might reduce the accuracy of fingerprint verification systems. In the systems that verify the identity of two fingerprints using fingerprint features, it is critical to extract the correct feature information. In order to deal with these problems, compensation of imperfect information can be performed using multiple snapshots of enrollee’s fingerprints. In this paper, additional fingerprint images are used in enrollment phase. Our experiments using FVC 2002 databases show that the enrollment using multiple snapshots improves the performance of the whole fingerprint verification system.

Local feature analysis with class information

In this paper, we propose a new feature extraction method for face recognition. This method is based on Local Feature Analysis (LFA), a local method for face recognition since it constructs kernels detecting local structures of a face. However, LFA has shown some problems for recognition due to the nature of unsupervised learning. Here, we point out the problems of LFA and propose a new feature extraction method with class information to overcome the shortcomings of LFA. Our method consists of three steps. First, using LFA, a set of local structures are extracted. Second, we select some extracted structures that are efficient for recognition. At last, we combine the selected local structures to represent them in a more compact form. This results in new bases which have compromised aspects between kernels of LFA and eigenfaces for face images. Throughout the experiments, our method has shown improvements on the face recognition over the previously proposed methods, LFA, eigenface, and fisherface.