Rank-1 Identification Rate Research Articles

Cross-modal group-relation optimization for visible–infrared person re-identification

Visible–infrared person re-identification (VIPR) plays an important role in intelligent transportation systems. Modal discrepancies between visible and infrared images seriously confuse person appearance discrimination, e.g., the similarity of the same class of different modalities is lower than the similarity between different classes of the same modality. Worse still, the modal discrepancies and appearance discrepancies are coupled with each other. The prevailing practice is to disentangle modal and appearance discrepancies, but it usually requires complex decoupling networks. In this paper, rather than disentanglement, we propose to measure and optimize modal discrepancies. We explore a cross-modal group-relation (CMGR) to describe the relationship between the same group of people in two different modalities. The CMGR has great potential in modal invariance because it considers more stable groups rather than individuals, so it is a good measurement for modal discrepancies. Furthermore, we design a group-relation correlation (GRC) loss function based on Pearson correlations to optimize CMGR, which can be easily integrated with the learning of VIPR’s appearance features. Consequently, our CMGR model serves as a pivotal constraint to minimize modal discrepancies, operating in a manner similar to a loss function. It is applied solely during the training phase, thereby obviating the need for any execution during the inference phase. Experimental results on two public datasets (i.e., RegDB and SYSU-MM01) demonstrate that our CMGR method is superior to state-of-the-art approaches. In particular, on the RegDB dataset, with the help of CMGR, the rank-1 identification rate has improved by more than 7% compared to the case of not using CMGR.

Pairwise difference relational distillation for object re-identification

Most relationship knowledge distillation methods individually optimize pairwise similarities to improve the accuracy performance of a lightweight student network. However, this optimization approach may not be optimal for object re-identification (Re-ID) which prioritizes ranking. This is because it does not guarantee consistent ranking results between a lightweight student network and a large teacher network. For that, we propose a novel method called pairwise difference relational distillation (PDRD) for object Re-ID. First, we theoretically prove that minimizing the difference relationship between pairwise similarities resulting from student and teacher networks ensures consistent ranking results between the two networks. Second, based on this theoretical foundation, we combine non-linear activation functions on pairwise similarity discrepancies to create a non-linear pairwise difference rational knowledge loss function, which enhances knowledge transfer. Extensive experiments on four public datasets demonstrate that our method achieves state-of-the-art performance. For example, on Market-1501, using ResNet18 as a lightweight student network, our method acquires a rank-1 identification rate of 93.62%.

Semi-supervised LDA pedestrian re-identification algorithm based on K-nearest neighbor resampling

Person re-identification identify a specific person in surveillance network by similarity measurement between images of different camera views. However, existing metric learning based methods suffer from over-fitting problem. To solve this problem, a resampled linear discriminant analysis (LDA) method was proposed based on the statistical and topological characteristics of pedestrian images. This method utilized the k-nearest neighbours to form potential positive sample pairs. The potential positive pairs are used to improve the metric model and generalize the metric model to the test data. By minimizing the inter-class divergence of potential positive sample pairs, a semi-supervised re-sampling LDA person re-identification algorithm was established. It was then tested on the VIPeR, CUHK01 and Market 1501datasets. The results show that the proposed method achieves the best performance compared to some available methods. Especially, the proposed method outplays the best comparison method by 0.6% and 5.76% at rank-1 identification rate on the VIPeR and CUHK01 datasets respectively. At the same time, the improved LDA algorithm has improved the rank-1 identification accuracy of traditional LDA method by 9.36% and 32.11% on these two datasets respectively. Besides, the proposed method is limited to Market-1501 dataset when the test data is of large size.

A Multistep Fusion Matcher Approach for Large Scale Latent Fingerprint/Palmprint Recognition

Latent fingerprints are ubiquitously used as forensic evidence by law enforcement agencies in solving crimes. However, due to deformations and artifacts within latent fingerprint images, performance of the automated latent recognition systems are far from desired levels. A basic matcher specifically designed for clean fingerprints using a minutiae-based matching algorithm can have high speed and accuracy in a sensor-to-sensor matching task, but low accuracy in matching latent prints, due to scale, rotation and quality differences between latent and sensor images. In this study, we propose a unique multistep fusion matcher (FM) on top of a base matcher that would utilize scale, rotation, and quality attributes of minutiae with speed, memory, and accuracy trade options in the latent recognition process. FM match characteristics are analyzed by using a private dataset consisting of 5560 latent and 1M slap/rolled fingerprint images. In addition, 292 domain expert selected latents are used to compare the nationwide performance of the proposed method. FM's with multiresolution fusion (MRF) option have achieved competitive accuracy rates when searching 292 latent against 1 million background and projecting predictions for 69 million background. On the NIST SD302 public dataset, FM6 (FM option prioritizing accuracy for latent-to-sensor search) with MRF correctly recognizes 911 latent in rank-1, while the COTS system referenced in the NIST SD302 documentation recognizes only 790 from a gallery composed of 5950 latent and 100K rolled background database. FM6 MRF rank-1 count for 10K latent of NIST SD302 is 1415, whereas NIST?s referenced matcher rank-1 count is 880 for the same dataset. In addition, NIST SD302 rank-1 latents are used to construct 722 latent pairs to evaluate latent-to-latent matching performance. FM8 (FM option prioritizing accuracy for latent-to-latent search) with MRF has 46.1% rank-1 identification rate for latent-to-latent search against 10K latent background. Moreover, on a private 1457 latent palmprint versus 2296 sensor palmprint background, a palm matcher designed by dividing latent and palm images into 512x512 pixel segments produces 85.45% rank-1 accuracy by using FM6.

Occlusion-Aware Human Mesh Model-Based Gait Recognition

Partial occlusion of the human body caused by obstacles or a limited camera field of view often occurs in surveillance videos, which affects the performance of gait recognition in practice. Existing methods for gait recognition against occlusion require a bounding box or the height of a full human body as a prerequisite, which is unobserved in occlusion scenarios. In this paper, we propose an occlusion-aware model-based gait recognition method that works directly on gait videos under occlusion without the above-mentioned prerequisite. Specifically, given a gait sequence that only contains non-occluded body parts in the images, we directly fit a skinned multi-person linear (SMPL)-based human mesh model to the input images without any pre-normalization or registration of the human body. We further use the pose and shape features extracted from the estimated SMPL model for recognition purposes, and use the extracted camera parameters in the occlusion attenuation module to reduce intra-subject variation in human model fitting caused by occlusion pattern differences. Experiments on occlusion samples simulated from the OU-MVLP dataset demonstrated the effectiveness of the proposed method, which outperformed state-of-the-art gait recognition methods by about 15% rank-1 identification rate and 2% equal error rate in the identification and verification scenarios, respectively.

Deep Learning-Based SOLO Architecture for Re-Identification of Single Persons by Locations

Analyzing and judging of captured and retrieved images of the targets from the surveillance video cameras for person re-identification have been a herculean task for computer vision that is worth further research. Hence, re-identification of single persons by locations based on single objects by locations (SOLO) model is proposed in this paper. To achieve the re-identification goal, we based the training of the re-identification model on synchronized stochastic gradient descent (SGD). SOLO is capable of exploiting the contextual cues and segmenting individual persons by their motions. The proposed approach consists of the following steps: (1) reformulating the person instance segmentation as: (a) prediction of category and (b) mask generation tasks for each person instance, (2) dividing the input person image into a uniform grids, i.e., G×G grid cells in such a way that a grid cell can predict the category of the semantic and masks of the person instances provided the center of the person falls into the grid cell and (3) conducting person segmentation. Discriminating features of individual persons are obtained by extraction using convolution neural networks. On person re-identification Market-1501 dataset, SOLO model achieved mAP of 84.1% and 93.8% rank-1 identification rate, higher than what is achieved by other comparative algorithms such as PL-Net, SegHAN, Siamese, GoogLeNet, and M 3 L (IBN-Net50). On person re-identification CUHK03 dataset, SOLO model achieved mAP of 82.1 % and 90.1% rank-1 identification rate, higher than what is achieved by other comparative algorithms such as PL-Net, SegHAN, Siamese, GoogLeNet, and M 3 L (IBN-Net50). These results show that SOLO model achieves best results for person re-identification, indicating high effectiveness of the model. The research contributions are: (1) Application of synchronized stochastic gradient descent (SGD) to SOLO training for person re-identification and (2) Single objects by locations using semantic category branch and instance mask branch instead of detect-then-segment method, thereby converting person instance segmentation into a solvable problem of single-shot classification.

Biometric Performance as a Function of Gallery Size

Many developers of biometric systems start with modest samples before general deployment. However, they are interested in how their systems will work with much larger samples. To assist them, we evaluated the effect of gallery size on biometric performance. Identification rates describe the performance of biometric identification, whereas ROC-based measures describe the performance of biometric authentication (verification). Therefore, we examined how increases in gallery size affected identification rates (i.e., Rank-1 Identification Rate, or Rank-1 IR) and ROC-based measures such as equal error rate (EER). We studied these phenomena with synthetic data as well as real data from a face recognition study. It is well known that the Rank-1 IR declines with increasing gallery size, and that the relationship is linear against log(gallery size). We have confirmed this with synthetic and real data. We have shown that this decline can be counteracted with the inclusion of additional information (features) for larger gallery sizes. We have also described the curves which can be used to predict how much additional information would be required to stabilize the Rank-1 IR as a function of gallery size. These equations are also linear in log(gallery size). We have also shown that the entire ROC-curve was not systematically affected by gallery size, and so ROC-based scalar performance metrics such as EER are also stable across gallery size. Unsurprisingly, as additional uncorrelated features are added to the model, EER decreases. We were interested in determining the impact of adding more features on the median, spread and shape of similarity score distributions. We present evidence that these decreases in EER are driven primarily by decreases in the spread of the impostor similarity score distribution.

Heterogeneous feature-aware Transformer-CNN coupling network for person re-identification.

Person re-identification plays an important role in the construction of the smart city. A reliable person re-identification system relieves users from the inefficient work of identifying the specific individual from enormous numbers of photos or videos captured by different surveillance devices. The most existing methods either focus on local discriminative features without global contextual information or scatter global features while ignoring the local features, resulting in ineffective attention to irregular pedestrian zones. In this article, a novel Transformer-CNN Coupling Network (TCCNet) is proposed to capture the fluctuant body region features in a heterogeneous feature-aware manner. We employ two bridging modules, the Low-level Feature Coupling Module (LFCM) and the High-level Feature Coupling Module (HFCM), to improve the complementary characteristics of the hybrid network. It is significantly helpful to enhance the capacity to distinguish between foreground and background features, thereby reducing the unfavorable impact of cluttered backgrounds on person re-identification. Furthermore, the duplicate loss for the two branches is employed to incorporate semantic information from distant preferences of the two branches into the resulting person representation. The experiments on two large-scale person re-identification benchmarks demonstrate that the proposed TCCNet achieves competitive results compared with several state-of-the-art approaches. The mean Average Precision (mAP) and Rank-1 identification rate on the MSMT17 dataset achieve 66.9% and 84.5%, respectively.

Learning directly from synthetic point clouds for “in-the-wild” 3D face recognition

Point clouds-based networks have achieved great attention in 3D object classification, segmentation, and indoor scene semantic parsing, but its application to 3D face recognition is still underdeveloped owing to two main reasons: lack of large-scale 3D facial data and absence of deep neural network that can directly extract discriminative face representations from point clouds. To address these two problems, a PointNet++ based network is proposed in this paper to extract face features directly from point clouds facial scans and a statistical 3D Morphable Model based 3D face synthesizing strategy is established to generate large-scale unreal facial scans to train the proposed network from scratch. A curvature-aware point sampling technique is proposed to hierarchically down-sample feature-sensitive points which are crucial to pass and aggregate discriminative facial features deeply. In addition, a novel 3D face transfer learning method is proposed to ease the domain discrepancy between synthetic and ‘in-the-wild’ faces. Experimental results on two public 3D face benchmarks show that the network trained only on synthesized data can also be well generalized to ‘in-the-wild’ 3D face recognition. Our method achieves the state-of-the-art results by achieving an overall rank-1 identification rate of 99.46% and 99.65% on FRGCv2 and Bosphorus, respectively. Further, we evaluate on a self-collected dataset to demonstrate the robustness and application potential of our method.

End-to-End Automated Latent Fingerprint Identification With Improved DCNN-FFT Enhancement.

Automatic Latent Fingerprint Identification Systems (AFIS) are most widely used by forensic experts in law enforcement and criminal investigations. One of the critical steps used in automatic latent fingerprint matching is to automatically extract reliable minutiae from fingerprint images. Hence, minutiae extraction is considered to be a very important step in AFIS. The performance of such systems relies heavily on the quality of the input fingerprint images. Most of the state-of-the-art AFIS failed to produce good matching results due to poor ridge patterns and the presence of background noise. To ensure the robustness of fingerprint matching against low quality latent fingerprint images, it is essential to include a good fingerprint enhancement algorithm before minutiae extraction and matching. In this paper, we have proposed an end-to-end fingerprint matching system to automatically enhance, extract minutiae, and produce matching results. To achieve this, we have proposed a method to automatically enhance the poor-quality fingerprint images using the “Automated Deep Convolutional Neural Network (DCNN)” and “Fast Fourier Transform (FFT)” filters. The Deep Convolutional Neural Network (DCNN) produces a frequency enhanced map from fingerprint domain knowledge. We propose an “FFT Enhancement” algorithm to enhance and extract the ridges from the frequency enhanced map. Minutiae from the enhanced ridges are automatically extracted using a proposed “Automated Latent Minutiae Extractor (ALME)”. Based on the extracted minutiae, the fingerprints are automatically aligned, and a matching score is calculated using a proposed “Frequency Enhanced Minutiae Matcher (FEMM)” algorithm. Experiments are conducted on FVC2002, FVC2004, and NIST SD27 latent fingerprint databases. The minutiae extraction results show significant improvement in precision, recall, and F1 scores. We obtained the highest Rank-1 identification rate of 100% for FVC2002/2004 and 84.5% for NIST SD27 fingerprint databases. The matching results reveal that the proposed system outperforms state-of-the-art systems.

An ensemble of fingerprint matching algorithms based on cylinder codes and mtriplets for latent fingerprint identification

Automatic latent fingerprint identification is beneficial during forensic investigations. Usually, latent fingerprint identification algorithms are used to find a subset of similar fingerprints from those previously captured on databases, which are finally examined by latent examiners. Yet, the identification rate achieved by latent fingerprint identification algorithms is far from those obtained by latent examiners. One approach for improving identification rates is the fusion of the match scores computed with fingerprint matching algorithms using a supervised classification algorithm. This approach fuses the results provided by different lower-level algorithms to improve them. Thus, we propose a fusion of fingerprint matching algorithms using a supervised classifier. Our proposal starts with two different local matching algorithms. We substitute their global matching algorithms with another independent of the local matching, creating two lower-level algorithms for fingerprint matching. Then, we combine the output of these lower-level algorithms using a supervised classifier. Our proposal achieves higher identification rates than each lower-level algorithm and their fusion using traditional approaches for most of the rank values and reference databases. Moreover, our fusion algorithm reaches a Rank-1 identification rate of $$74.03\%$$ and $$71.32\%$$ matching the 258 samples in the NIST SD27 database against 29,257 and 100,000 references, the two largest reference databases employed in our experiments.

CNNAI: A Convolution Neural Network-Based Latent Fingerprint Matching Using the Combination of Nearest Neighbor Arrangement Indexing.

Automatic fingerprint identification systems (AFIS) make use of global fingerprint information like ridge flow, ridge frequency, and delta or core points for fingerprint alignment, before performing matching. In latent fingerprints, the ridges will be smudged and delta or core points may not be available. It becomes difficult to pre-align fingerprints with such partial fingerprint information. Further, global features are not robust against fingerprint deformations; rotation, scale, and fingerprint matching using global features pose more challenges. We have developed a local minutia-based convolution neural network (CNN) matching model called “Combination of Nearest Neighbor Arrangement Indexing (CNNAI).” This model makes use of a set of “n” local nearest minutiae neighbor features and generates rotation-scale invariant feature vectors. Our proposed system doesn't depend upon any fingerprint alignment information. In large fingerprint databases, it becomes very difficult to query every fingerprint against every other fingerprint in the database. To address this issue, we make use of hash indexing to reduce the number of retrievals. We have used a residual learning-based CNN model to enhance and extract the minutiae features. Matching was done on FVC2004 and NIST SD27 latent fingerprint databases against 640 and 3,758 gallery fingerprint images, respectively. We obtained a Rank-1 identification rate of 80% for FVC2004 fingerprints and 84.5% for NIST SD27 latent fingerprint databases. The experimental results show improvement in the Rank-1 identification rate compared to the state-of-art algorithms, and the results reveal that the system is robust against rotation and scale.

A Comparative Performance Analysis of Multimodal-Multialgorithm System Framework Based on Rank Level Fusion

The Unimodal biometric framework have various fundamental issues, for example, intra-class alteration, noisy data, failure-to-enroll, spoofing attacks, unacceptable error rate and non-universality. To defeat this shortcoming multibiometric is a decent alternative where we can utilize at least two individual modalities. This paper gives a comparative analysis of multi-algorithm and multimodal system framework based on rank level fusion. An effective combination strategy that integrates information given by different domain specialist dependent on rank level fusion approach is utilized to enhance the presentation of the framework. The rank of individual matcher is combined using the highest rank, Borda count, weighted Borda count, nonlinear weighted approach and Bucklin combination approach. The outcomes of the results show there is a noteworthy exhibition enhancement in the identification accuracy can be accomplished when contrasted those from unimodal frameworks. The outcomes also reveal that combination of individual modalities can enhance the biometric system performance. The experiment based on multimodal (NIST BSSR1 multimodal database of fingerprint and face) and multialgorithm (Hong Kong Polytechnic University database of palmprint) system shows an improvement in term of the Rank-1 identification rate of the system.

Semisupervised Consistent Projection Metric Learning for Person Reidentification.

Person reidentification is a hot topic in the computer vision field. Many efforts have been paid on modeling a discriminative distance metric. However, existing metric-learning-based methods are a lack of generalization. In this article, the poor generalization of the metric model is argued as the biased estimation problem that the independent identical distribution hypothesis is not valid. The verification experimental result shows that there is a sharp difference between the training and test samples in the metric subspace. A semisupervised consistent projection metric-learning method is proposed to ease the biased estimation problem by learning a consistent constrained metric subspace in which the identified pairs are forced to follow the distribution of the positive training pairs. First, a semisupervised method is proposed to generate potential matching pairs from the k -nearest neighbors of test samples. The potential matching pairs are used to estimate the distances' distribution center of the positive test pairs. Second, the metric subspace is improved by forcing this estimation to be close to the center of the positive training pairs. Finally, extensive experiments are conducted on five datasets and the results demonstrate that the proposed method reaches the best performance, especially on the rank-1 identification rate.

Biometric Authentication through Unification of Finger Dorsal Biometric Traits

This paper presents a novel method for personal authentication of individuals using a deep learning approach. The reported method implements the fusion of two biometric attributes, namely fingernail plates and finger knuckles. Both these biometric traits have been extracted from the index, middle and ring fingers of a database of 890 (5 each from 178 persons) low-resolution hand dorsal images. AlexNet, a pre-trained deep learning model, has been tailored for feature extraction to suit the number of classes in the database. Deep learning features are extracted from both the traits using the customized Convolutional Neural Network (CNN). These features have been fused at the score-level and the rank-level for different combinations. Both these levels of fusion have been executed through well-established rules. Score-level fusion results show that verification can be done quite accurately with this multimodal framework. Results also depict that information fusion of these traits at the rank-level can be optimized to give Rank-1 identification rate of a perfect 100%.

Large-Scale Person Re-Identification Based on Deep Hash Learning.

Person re-identification in the image processing domain has been a challenging research topic due to the influence of pedestrian posture, background, lighting, and other factors. In this paper, the method of harsh learning is applied in person re-identification, and we propose a person re-identification method based on deep hash learning. By improving the conventional method, the method proposed in this paper uses an easy-to-optimize shallow convolutional neural network to learn the inherent implicit relationship of the image and then extracts the deep features of the image. Then, a hash layer with three-step calculation is incorporated in the fully connected layer of the network. The hash function is learned and mapped into a hash code through the connection between the network layers. The generation of the hash code satisfies the requirements that minimize the error of the sum of quantization loss and Softmax regression cross-entropy loss, which achieve the end-to-end generation of hash code in the network. After obtaining the hash code through the network, the distance between the pedestrian image hash code to be retrieved and the pedestrian image hash code library is calculated to implement the person re-identification. Experiments conducted on multiple standard datasets show that our deep hashing network achieves the comparable performances and outperforms other hashing methods with large margins on Rank-1 and mAP value identification rates in pedestrian re-identification. Besides, our method is predominant in the efficiency of training and retrieval in contrast to other pedestrian re-identification algorithms.

Multi-view 3D face reconstruction with deep recurrent neural networks

Image-based 3D face reconstruction has great potential in different areas, such as facial recognition, facial analysis, and facial animation. Due to the variations in image quality, single-image-based 3D face reconstruction might not be sufficient to accurately reconstruct a 3D face. To overcome this limitation, multi-view 3D face reconstruction uses multiple images of the same subject and aggregates complementary information for better accuracy. Though appealing, there are multiple challenges in practice. Among these challenges, the most significant is the difficulty to establish coherent and accurate correspondence among a set of images, especially when these images are captured under unconstrained in-the-wild condition. This work proposes a method, Deep Recurrent 3D FAce Reconstruction (DRFAR), to solve the task of multi-view 3D face reconstruction using a subspace representation of the 3D facial shape and a deep recurrent neural network that consists of both a deep convolutional neural network (DCNN) and a recurrent neural network (RNN). The DCNN disentangles the facial identity and the facial expression components for each single image independently, while the RNN fuses identity-related features from the DCNN and aggregates the identity specific contextual information, or the identity signal, from the whole set of images to estimate the facial identity parameter, which is robust to variations in image quality and is consistent over the whole set of images. Experimental results indicate significant improvement over state-of-the-art in both the accuracy and the consistency of 3D face reconstruction. Moreover, face recognition results on IJB-A with the UR2D face recognition pipeline indicate that, compared to single-view 3D face reconstruction, the proposed multi-view 3D face reconstruction algorithm can improve the face identification accuracy of UR2D by two percentage points in Rank-1 identification rate.

Geometric statistics-based descriptor for 3D ear recognition

Several feature keypoint detection and description techniques have been proposed in the literature for 3D shape recognition. These techniques work well in discriminating different classes of shapes; however, they fail when used for comparing highly similar objects such as 3D ear or face in biometric applications. In this paper, we propose an efficient feature keypoint detection and description technique using geometric statistics for representation and matching of highly similar 3D objects and demonstrate its effectiveness in 3D ear-based biometric recognition. To compute the descriptor, we first extract feature keypoints from the 3D data by making use of surface variations followed by defining a descriptor vector for each keypoint. The descriptor vector is generated using three components. To compute the first component, concentric spheres that divide the space around a keypoint into annular regions are considered. Points falling in the annular regions are projected onto a plane perpendicular to the oriented normal of the keypoint. Lower-order moments of the 2D histogram of the spatial distribution of these projected points for each annular region are computed and used to define the first component of the descriptor vector. Next, component of the descriptor vector is computed using histograms of the inner products of the normals of the keypoint and its neighbours. The third component of the descriptor vector encodes the signed distances of the neighbours of the keypoint from the projection plane. Before concatenating individual components of the descriptor vector, the values are normalized to a common scale. Experiments on University of Notre Dame public database-Collection J2 (UND-J2) have achieved a rank-1 and rank-2 identification rates of $$98.60\%$$ and $$100\%$$, respectively, with an equal error rate of $$1.50\%$$. Comparative results show the superiority of the proposed descriptor in recognizing highly similar objects like human ear.

Single-sensor hand-vein multimodal biometric recognition using multiscale deep pyramidal approach

Biometrics has emerged as a powerful technology for person authentication in various scenarios including forensic and civilian applications. Deployment of biometric solutions that use cues from multiple modalities enhances the reliability and robustness of authentication necessary to meet the increasingly stringent security requirements. However, there are two drawbacks typically associated with multimodal biometrics. Firstly, the image acquisition process in such systems is not very user-friendly, primarily due to the time and effort required to capture biometric samples belonging to multiple modalities. Secondly, the overall cost is higher as they employ multiple biometric sensors. To overcome these drawbacks, we employ a single NIR sensor-based image acquisition in the proposed approach for hand-vein recognition. From the input hand image, a palm-vein and four finger-vein subimages are extracted. These images are then enhanced by CLAHE and transformed into illumination invariant representation using center-symmetric local binary pattern (CS-LBP). Further, a hierarchical non-rigid matching technique inspired by the architecture of deep convolutional networks is employed for matching the CS-LBP features. Finally, weighted sum rule-based matching score-level fusion is performed to combine the palm-vein and the four finger-vein modalities. A set of rigorous experiments has been performed on an in-house database collected from the left and right hands of 185 subjects and the publicly available CASIA dataset. The proposed approach achieves equal error rates of 0.13% and 1.21%, and rank-1 identification rates of 100% and 100% on the in-house and CASIA datasets, respectively. Additionally, we compare the proposed approach with the state-of-the-art techniques proposed for vascular biometric recognition in the literature. The important findings are (1) the proposed approach outperforms all the existing techniques considered in this study, (2) the fusion of palm-vein and finger-vein modalities consistently leads to better performance for all the feature extraction techniques considered in this work. (3) Furthermore, our experimental results also suggest that considering the constituent palm-vein and finger-vein images instead of the entire hand-vein images achieves better performance.

Biometric recognition via texture features of eye movement trajectories in a visual searching task.

Biometric recognition technology based on eye-movement dynamics has been in development for more than ten years. Different visual tasks, feature extraction and feature recognition methods are proposed to improve the performance of eye movement biometric system. However, the correct identification and verification rates, especially in long-term experiments, as well as the effects of visual tasks and eye trackers’ temporal and spatial resolution are still the foremost considerations in eye movement biometrics. With a focus on these issues, we proposed a new visual searching task for eye movement data collection and a new class of eye movement features for biometric recognition. In order to demonstrate the improvement of this visual searching task being used in eye movement biometrics, three other eye movement feature extraction methods were also tested on our eye movement datasets. Compared with the original results, all three methods yielded better results as expected. In addition, the biometric performance of these four feature extraction methods was also compared using the equal error rate (EER) and Rank-1 identification rate (Rank-1 IR), and the texture features introduced in this paper were ultimately shown to offer some advantages with regard to long-term stability and robustness over time and spatial precision. Finally, the results of different combinations of these methods with a score-level fusion method indicated that multi-biometric methods perform better in most cases.