Figure 7 - uploaded by Venu Govindaraju
Content may be subject to copyright.
Illustration of two fingerprints of the same user with marked minutiae and the corresponding adjacency graph based on the K-plet representaion. In this particular example each minutia is connected to its K=4 nearest neighbors. It is also to be noted that the topologies of the graphs are different due to an extra unmatched minutiae in the left print
Source publication
In this paper, we present a new fingerprint matching algorith m based on graph matching principles. We define a new representation called K-plet to encode the local neighborhood of each minutia. We also present CBFS (Coupled BFS), a new dual graph traversal algorithm for consolidating all the local neighborhood matches and analyze its computational...
Context in source publication
Context 1
... encode the local structural relationship of the K-plet formally in the form of a graph G(V, E). Each minutia is represented by a vertex v and each neighboring minutiae is represented by a directed edge (u, v) (See Figure 7). Each vertex u is colored with attributes (xu, yu, θu, tu) that represents the co-ordinate, orientation and type of minutiae(ridge ending or bifurcation). ...
Citations
... The images have a resolution of 2040 × 2040 pixels at 500 ppi with 256 grey levels. In order to test the potential of MSA in improving matching accuracy, we have implemented -nearest neighbor matching (Chikkerur, Cartwright, & Govindaraju, 2005) implemented by Ghafoor et al. (2020) on the minutiae shortlisted by MSA. In the -nearest method, each minutia is encoded based on its nearest neighboring minutiae. ...
... But for a challenging dataset like THUPALMLAB (2012), very tight thresholds can deteriorate matching accuracy. Since we have used -nearest neighbor matching method (Chikkerur et al., 2005) implemented by Ghafoor et al. (2020), the efficacy of offline application of MSA can be clearly established by comparing the performance of the matching algorithm with and without MSA in Table 5. Table 5 shows that the same matching algorithm is able to perform better by choosing fewer minutiae using MSA. ...
Palmprints have gained enormous popularity in the biometric industry recently. Due to the large amount
of information provided by the palmprints and their forensic value, many large-scale identification systems
are employing high resolution palmprint-based biometric systems with minutiae as the primary identification
feature. Minutiae are the palm ridge endings or bifurcations that are represented by their location and
orientation, i.e., (𝑥, 𝑦, 𝜃). Due to the large size of palmprints, approximately 1000 minutiae are extracted
per palmprint on average. Besides the large number, another problem with the extracted minutiae is that
a considerable amount is false. Besides increasing the number of minutiae matches between palmprints, these
false minutiae reduce matching accuracy. In order to reduce minutiae matches and eliminate false minutia,
previous studies have adopted multifaceted approaches such as associating a quality or confidence factor with
each minutia, image registration, or inventing computationally efficient minutia descriptors. However, all these
approaches require associating each minutia with additional parameters. In this paper, we propose a simple
and intuitive histogram-based minutiae selection algorithm (MSA) using only the basic properties, i.e., (𝑥, 𝑦, 𝜃)
to shortlist a subset of best minutiae candidates for matching. This provides the dual benefit of: (1) a reduced
number of minutiae matches between palmprints, and (2) improved matching accuracy through the elimination
of false minutiae. The proposed method does not require estimating any additional parameters for minutiae
and is independent of minutia encoding and matching methods used. Our results are acquired on a popular
and challenging high resolution palmprint dataset and show the efficacy of the proposed algorithm through
a clear improvement in matching accuracy and a significant reduction in the number of required minutiae
matches between palmprints.
... The first category relies on relationships between the observed minutia and its neighboring minutiae in a fixed radius, i.e., star structure [13] and minutiae cylinder-code [14]. On the other hand, the second category forms the relationships between the observed minutia and the -nearest neighbors [15,16,17,18,19]. In this work, we focus on improving algorithms in the second category. ...
... Medina-Pérez et al. [18] increase the number of minutiae-triplets by selecting more than two nearest neighboring minutiae ( ) to provide matching redundancy. Chikkerur et al. [19] choose -nearest neighbors ( ) to build -plets representation. These algorithms increase the number of comparisons from redundant features and require more memory to store them. ...
Most fingerprint matching methods suffer from elastic deformation of fingerprints, resulting in an increment of the false-rejection rate. We propose a new distance model for the minutiae-triplets formation that can remedy the elastic deformation of fingerprints. The new distance model, called a directionally weighted distance model, provides higher priority to neighboring minutiae within the same direction or the same ridge flow of the observed minutia. We introduce two methods that apply the proposed distance model to the minutiae-triplets formation. While the first method directly applies the model, the second method combines the model with ridge flow to handle highly curved areas such as singular-point areas or highly distorted fingerprint areas. We evaluate the proposed methods using two minutiae-triplets matching algorithms on sixteen public domain fingerprint databases. The experimental results show that the proposed distance model can significantly improve the accuracy of both matching algorithms on most databases.
... Then, this set is queried with graph matching algorithms to Ąnd a correspondence with the input feature of the person to be identiĄed. Among these digital features, we Ąnd Ąngerprints [22,161], retina [71] and the face [140]. In computer vision, graph matching is used for image analysis and image databases for both indexing and retrieval [30,2,19,19,64]. ...
Graph Pattern Matching (GPM), usually evaluated through subgraph isomorphism, finds subgraphs of a large data graph that are similar to an input query graph. It has many applications, such as pattern recognition and finding communities in social networks.
However, besides its NP-completeness, the strict constraints of subgraph isomorphism are making it impractical for GPM in the context of big data. As a result, relaxed GPM models such as graph simulation emerged as they yield interesting results in polynomial time. Moreover, massive graphs generated by mostly social networks require distributed storing and processing of the data over multiple machines. Therefore, the existing algorithms for relaxed GPM need to be revised to this context by adopting new paradigms for big graph processing, e.g. Think-Like-A-Vertex and its derivatives.
In this thesis, we investigate the use of distributed graph processing paradigms and systems in the evaluation of GPM queries. Our goal is to identify the programming models that are best suited for this problem. Furthermore, we study the existing GPM approaches, with more emphasis on the relaxed ones in the aim of proposing new parallel and distributed algorithms for relaxed GPM that guarantee linear scalability.
Our contributions are summarized as follows. First, we propose a taxonomy of prior work on distributed GPM based on multiple criteria, such as the GPM model and the programming paradigm. Next, we introduce BDSim as a new model that captures more semantic similarities compared to the existing models while being feasible in cubic time. Besides, we design distributed vertex-centric algorithms that are adapted to the context of massive graphs for evaluating BDSim. Furthermore, we propose the first fully distributed and scalable approach for strong simulation, a relaxed GPM model that strikes a balance between flexibility and tractability. Finally, we propose the first efficient parallel edge-centric approach for evaluating graph simulation and dual simulation in distributed graphs. We validate the effectiveness and efficiency of our approaches through theoretical guarantees and reliable testing over synthetic and real-world graphs.
We confirmed in this thesis that different paradigms can be used in designing distributed GPM algorithms depending on the GPM model adopted. Indeed, algorithms for neighborhood-based models such as subgraph isomorphism and strong simulation perform better with a vertex-centric or subgraph-centric paradigm as the latter involves some data locality, while the most efficient algorithms for graph simulation and dual simulation are edge-based and offer linear scalability guarantees.
... Therefore, for performance, additional functions such as kmeans clustering must be used in quadruplet-based schemes, which requires more computational complexity [5]. A k-pletbased representation of minutia has been first proposed in [6]. It consists of k-plet and all other minutiae within its fixed radius. ...
A simple fingerprint identification scheme compares an input fingerprint with all the fingerprints in the database to find any matching fingerprint. That is, the simple matching method considers all fingerprints in the database as candidates for a given input fingerprint. However, this simple matching method requires a lot of processing time. To reduce the processing time, we can use fingerprint indexing to reduce the size of a candidate set for an input fingerprint. The candidate set is the set of fingerprints most similar to the input fingerprint. Usually, the size of the candidate set is much smaller than the size of the whole fingerprint database. It enables efficient identification by comparing the input fingerprint with only the fingerprints in the candidate set instead of the entire database. In this paper, we analyze the index distribution of the Kavati et al.’s indexing method and propose a new fingerprint index vector which tries to make the index distribution more similar to the uniform distribution. Our new index vector consists of elements that are not highly correlated, which is measured by the Pearson correlation coefficient. Because our indexing method makes the index values widely spread over the index space, it reduces the number of candidates for a given fingerprint in fingerprint identification. Our indexing method shows a higher match rate with a smaller candidate set than the existing triplet-based indexing methods. Especially, our indexing method is up to 6.4 times more accurate than the Kavati et al.’s indexing method. Our result shows that the index distribution significantly affects performance of indexing methods.
... The overall palmprint identification system is divided into three phases: (i) palmprint preprocessing and enhancement; (ii) palmprint feature extraction and encoding; and (iii) palmprint matching [10,11]. The performance and accuracy of the palmprint identification system heavily depend on the quality of the extracted minutia features. ...
... Not only are minutiae-based algorithms accurate, but they also support latent (partial) palmprint matching, which is very useful in criminal and forensic applications. The basic methodology of the most robust palmprint encoding and matching methods follows the same technique where each minutia is taken as a reference, and it is encoded based on its distances and angular differences with the neighbouring minutiae [10,11]. Tico and Kuosmanen [20] proposed a technique where the orientation fields around individual minutiae are considered for encoding and matching. ...
... The parameters at each stage of matching are empirically chosen to minimise the EER. Table 6 reports the EER comparisons of the proposed algorithm with open-source implementations of Chikkerur's K-plet [11] and NIST Bozorth3 [32] matching techniques, and also with other works published in the literature. It can be observed that the proposed two-stage matching algorithm significantly outperforms both K-plet [11] and Bozorth3 [32], and other popular matching algorithms on THUPALMLAB palmprint database. ...
Palmprint‐based human authentication has shown great potential for civil, forensic, and corporate security applications in recent years. Palmprint recognition systems suffer because of large palmprint sizes and the presence of a large number of creases and erroneous minutiae that make the enhancement and matching phases a challenge. In this study, a novel approach is presented based on efficient enhancement and a two‐stage matching technique that demonstrates highly accurate identification results. The enhancement approach extracts minutia features from high‐quality regions based on local ridge characteristics. The selected minutiae are then matched using a two‐stage local and global minutiae neighbour‐based matching technique. To demonstrate the performance of the proposed technique, comparisons with open‐source algorithms are made based on equal error rate and detection error trade‐off graph. The results confirm the efficacy of proposed palmprint enhancement and identification technique.
... Local minutiae descriptors [8,11,13,18,20,21] are used for alignment and a global linking is used for matching. The alignment process is not very accurate in poor quality images [22]. k-plet-graph matching [22]. ...
... The alignment process is not very accurate in poor quality images [22]. k-plet-graph matching [22]. No need of explicit alignment of minutiae. ...
... The majority of the existing minutiae-based methods used single minutiae representation. This adopts one-to-one correspondence with minutiae coordinates and direction [22], which is more sensitive to geometric transformations and needs an additional image alignment process before matching. In minutiae triplets, Kovacs-Vajna [42] used a connected graph composed of triangles with shared edges with the features of three angles and three distances to overcome deformation. ...
An invariant algorithm for matching fingerprints or palmprints is presented for hand biometrics where only parts of fingerprints or palmprints may be available in images for comparison. The algorithm uses an extended version of the minutiae-based approach, treating the pattern as a graph of minutiae points. A set of such points is subjected to Delaunay triangulation yielding a starting set of base-triangles for matching. There can be multiple matches of such triangles between the template and test-as similar triangles with a tolerance in the angles. Graphs are then grown to five or more nodes as long as a match can be found, until the maximum size matching graph is obtained. If the test matches a significant part of the template, the maximum order of graph matched will be high. The matching process is robust to transformations such as rotation, translation and scale changes. It can be applied to any part of the hand provided minutiae-like points are identifiable prior to the matching steps. The algorithm is tested using a set of fingerprint images from FVC 2002 DB1 with geometrical variations. One hundred genuine and 5048 impostor scores are generated from a set of images with sufficient geometrical variations, which yields an EER of about 6%. Validation is performed with a set of partial quality palmprint segments from the THUPALMLAB database. Among 70 genuine and 630 imposter scores, an EER of 4.71% is reported. The experiment was extended using partially and fully degraded image segments with a arnge of geometrical variations. Among 1526 genuine and 59296 impostor scores, an EER of 6.31% is reported. It proves the principle behind the methodology and demonstrates that the method can be effective with degraded fingerprint and palmprint image segments and is robust to similarity transformations present in the data. In addition, it can be applied for forensic matching from the palm or any parts from the hand. 1 Introduction Fingerprints are extensively used in many applications such as access control, transactions, border-crossing, e-passport, criminal analysis and bio-cryptography whereas palmprints are mostly used in crime-scenes analysis and access control. For automated identification , fingerprint image acquisition is processed with a contact sensor where a finger directly touches the surface of the device, or using a contactless method with a high resolution digital camera at a distance. The majority of the images in access control systems are captured by touch-based sensors using fingertip or palmprint with low-resolution, whereas the images in forensic applications are captured as high resolution latent prints. At present, there is no existing system which can adopt both fingerprints and palmprints in hand bio-metrics for both access control and forensic applications. This work explores a common algorithm for hand prints to apply in a forensic scenario in future by reducing manual comparisions of two prints and extending it to civilain applications as well. Generally, touch-based palm print matching is categorized based on the source and size of the images such as: i) full-to-full palm print; ii) latent-to-full palm print; and iii) live-scan partial-to-full palm print. Among these types of contact-based palmprint matching, full-to-full palmprint matching is still not used in many applications as it needs a high level of accuracy. However, latent-to-full palmprint matching techniques are still a challenge in forensic applications due to the lower quality of images, nonlinear image distortion and large number of spurious minutiae. Moreover, matching is performed by handling manual alignment or manual marking of some key points. The application of partial-to-full palm print matching is used in access-control which has the requirement to register or to be able to determine which region of the hand the partial print is obtained from, and this remains a challenge and requires further investigation and improvement. On the other hand, fingerprints have been an attractive biomet-ric modality, especially for forensic applications for over the years. Automated fingerprint identification and retrieval work well when the acquisition systems during enrollment and test are matched and the images are of good quality. However, for forensics and law enforcement, latent prints (photographic images transferred from crime scenes) have to be matched with entities in a database of fingerprints and palm prints acquired using other sensors and this is still a difficult proposition. Latent prints are often of poor quality and degraded in parts. They are unable to be pre-processed successfully enough for automated minutiae matching to perform well. High-resolution cameras make it possible to obtain detailed images of palms and fingers and they have been investigated as potential biometric modalities. However, photographic images have problems of geometric variations with varying illumination, and poor quality from being outside the depth of focus. In comparision with fingerprints , palm prints produce a large number of minutiae (x10 times) and non-linear distortions would be a challenging issue due to the hand structure and its increased surface area. Therefore, conventional fingerprint matching methods are not directly suitable for such palm prints. However, ridge features and some fiducial points similar to minutiae can be identified in high-resolution images of palms and fingers. If an automated method of matching can use minutiae and other such points for matching images that exhibit large variations in quality through geometrical transformations or partial occlusion as shown in Figure 1, it can narrow down the set of prints retrieved from a database that will need to be manually examined. Within this context, we have investigated an alternate methodology of matching that is suitable for both fingerprint and palmprint in a common platform. As this work explores a common algorithm for hand prints to be applied in forensic scenario by reducing manual comparisions of two prints, it aims to investigate the approach that can be used for both
... The minutiae are detected from the thinned image that is morphologically computed from the enhanced fingerprint image [18], [19]. Chikkerur et al. [20] proposed a K-plet structure-based local minutiae encoding technique and performed minutiae matching based on a dual-graph traversal algorithm. NNs were employed by Jea and Govindaraju [21] to compute a normalized matching score in overlapping regions of template and candidate fingerprints. ...
... The proposed fingerprint matching algorithm is compared with the open-source implementation of Chikkerur et al.'s K-plet [20] matching algorithm. The number of nearest neighbors for minutiae encoding is empirically chosen as 10 and the match threshold is set to 50% of encoded neighbors. ...
... The number of nearest neighbors for minutiae encoding is empirically chosen as 10 and the match threshold is set to 50% of encoded neighbors. The results are presented in Table IV, which shows that the proposed matching algorithm performs better than the K-plet [20] matching algorithm. ...
This article presents an efficient fingerprint identification system that implements an initial classification for search-space reduction followed by minutiae neighbor-based feature encoding and matching. The current state-of-the-art fingerprint classification methods use a deep convolutional neural network (DCNN) to assign confidence for the classification prediction, and based on this prediction, the input fingerprint is matched with only the subset of the database that belongs to the predicted class. It can be observed for the DCNNs that as the architectures deepen, the farthest layers of the network learn more abstract information from the input images that result in higher prediction accuracies. However, the downside is that the DCNNs are data hungry and require lots of annotated (labeled) data to learn generalized network parameters for deeper layers. In this article, a shallow multifeature view CNN (SMV-CNN) fingerprint classifier is proposed that extracts: 1) fine-grained features from the input image and 2) abstract features from explicitly derived representations obtained from the input image. The multifeature views are fed to a fully connected neural network (NN) to compute a global classification prediction. The classification results show that the SMV-CNN demonstrated an improvement of 2.8% when compared to baseline CNN consisting of a single grayscale view on an open-source database. Moreover, in comparison with the state-of-the-art residual network (ResNet-50) image classification model, the proposed method performs comparably while being less complex and more efficient during training. The result of classification-based fingerprint identification has shown that the search space is reduced by over 50% without degradation of identification accuracies.
... Global and local matching [6] are two ways of minutiae matching in the literature where local approach is attempted based on the structures of i) nearest neighbour; ii) fixed radius; and iii) minutiae triangle. Since local matching mostly uses the features of minutiae coordinates and minutiae direction [7], the one to one correspondence matching cannot be successful as transformation can be possible between the matching image pair and therefore, need a pre-alignment process before the matching. To overcome the issue, global matching are adopted in addition to local matching with minutiae related characteristics in few of the past works [8]- [10]. ...
A new algorithm for matching finger or palm prints is presented for use where the full hand is considered as a biometric and only parts may be available in images for comparison. The algorithm uses an extended version of the minutiae-based approach treating the pattern as a graph of minutiae-like points. The procedure to identify minutiae-like points uses Gabor filtering, edge detection and thinning and following line patterns. A set of such points is subjected to Delaunay triangulation yielding a starting set of base-triangles for matching. There can be multiple matches of such triangles between the template and test - as similar triangles with a tolerance in the angles. Graphs are then grown to 5 and more nodes as long as a match can be found, until the maximum size matching graph is obtained. If the test matches a significant part of the template, the maximum order of graph matched will be high. The matching process is robust to transformations such as rotation, translation and scale changes. It can be applied to any part of the hand provided minutiae-like points are identifiable prior to the matching steps. The algorithm is tested using 158 fingerprint images from FVC 2002 DB1. 100 genuine and 5048 impostor scores are generated from 46 templates and 112 testing images. It had an EER of about 6%. It proves the principle behind the methodology and demonstrates that the method can be effective with degraded fingerprint images and is robust to similarity transformations present in the data. It can be applied for forensic fingerprint matching from the palm or parts other than the fingertips. By using multiple parts and multiple templates, the accuracy of the method will be improved with fusion in future versions of the algorithm.
... Bozorth3 was published by NIST with its NIST Biometric Image Software (NBIS) package. It's a widely embraced fingerprint matching algorithm in research (see, e.g., [22,44]). ...
Biometric authentication is increasingly being used for large scale human authentication and identification, creating the risk of leaking the biometric secrets of millions of users in the case of database compromise. Powerful "fuzzy" cryptographic techniques for biometric template protection, such as secure sketches, could help in principle, but go unused in practice. This is because they would require new biometric matching algorithms with potentially much diminished accuracy. We introduce a new primitive called a multisketch that generalizes secure sketches. Multisketches can work with existing biometric matching algorithms to generate strong cryptographic keys from biometric data reliably. A multisketch works on a biometric database containing multiple biometrics --- e.g., multiple fingerprints --- of a moderately large population of users (say, thousands). It conceals the correspondence between users and their biometric templates, preventing an attacker from learning the biometric data of a user in the advent of a breach, but enabling derivation of user-specific secret keys upon successful user authentication. We design a multisketch over tenprints --- fingerprints of ten fingers --- called TenSketch. We report on a prototype implementation of TenSketch, showing its feasibility in practice. We explore several possible attacks against TenSketch database and show, via simulations with real tenprint datasets, that an attacker must perform a large amount of computation to learn any meaningful information from a stolen TenSketch database.
... Although the encoding and matching techniques are different, yet the basic methodology remains the same. In order to encode a minutia, it is taken as a reference and its distance and angular difference with neighboring minutia points is computed and encoded [13][14][15]. Duta and Jain [16], extracted feature points along the prominent lines of palmprint images and performed matching by aligning the feature point pair patterns using a non-linear deformation model. Jain and Feng [8] proposed a minutia based encoding and matching technique for latent palmprint matching. ...
Automated human authentication is becoming increasingly important in today’s world due to increased need of security and surveillance applications deployed in almost all premises and installations. In this regard, palmprint biometric based identification has gained a lot of attention in recent years. However, due to large size of palmprint images and presence of principal lines, wrinkles, creases, and other noises, there are large number of inaccurate minutiae present. The computational requirement of palmprint identification is also quite large and it takes a lot of time to find identity of a palmprint in large database. In this study, a novel palmprint identification solution has been proposed that increases the accuracy of minutia detection based on improved frequency estimation and a novel region-quality based minutia extraction algorithm. Furthermore, a novel, efficient and highly accurate minutiae based encoding and matching algorithm is proposed that is designed to achieve maximum parallelism, and it is further accelerated using graphical processing unit. The results of the proposed palmprint identification demonstrate high accuracy and much faster identification speeds in comparison with current state of the art. Therefore, it can be considered as a robust, efficient and practical solution for palmprint based identification systems.
