Search

The viability of utilising a convolutional neural network-based (CNN-based) feature extractor together with a support vector machine (SVM) for the purpose of identity verification by means of near infra-red (NIR) images of individuals’ dorsal hand vein patterns is investigated in this paper. More specifically, this study aims to determine whether the utilisation of an SVM, instead of a typical softmax classifier, may lead to an increase in system performance within the context of hand vein-based authentication using CNNs. The proficiency of a variety of novel hand vein-based authentication systems is first gauged by employing a softmax classifier, after which the most proficient system is selected, retrained and re-evaluated with a SVM instead of a softmax classifier. The most proficient system, in which the softmax classifier is replaced with a SVM, achieves an accuracy of 98.90% and 99.23% respectively within the context of the Bosphorus and the Wilches datasets. Keywords: biometric authentication, hand vein, deep learning, similarity measure network, siamese networks, two- channel networks, segmentation, individual specific, convolutional neural networks, support vector machine. Title: Hand vein-based biometric authentication with convolutional neural networks and support vector machines Author: Emile Beukes, Johannes Coetzer International Journal of Recent Research in Electrical and Electronics Engineering (IJRREEE) ISSN 2349-7815 Vol. 9, Issue 3, July 2022 - September 2022 Page No: 12-31 Paper Publications Website: www.paperpublications.org Published Date: 04-August-2022 DOI: https://doi.org/10.5281/zenodo.6961864 Paper Download Link (Source) https://www.paperpublications.org/upload/book/Hand%20vein-based%20biometric-04082022-1.pdf, International Journal of Recent Research in Electrical and Electronics Engineering (IJRREEE), ISSN 2349-7815, Paper Publications, Website: www.paperpublications.org, {"references":["[1] \tN. A. Al-johania and L. A. Elrefaei, \"Dorsal hand vein recognition by convolutional neural networks: Feature learning and transfer learning approaches,\" International Journal of Intelligent Engineering and Systems, vol. 12, pp. 178–191, 2019.","[2] \tP. Arora, S. Srivastava, M. Hanmandlu, and S. Bhargava, \"Robust authentication using dorsal hand vein images,\" IEEE Intelligent Systems,, vol. 34, pp. 25–35, 2019.","[3] \tBeukes, \"Hand vein-based biometric authentication with limited training samples,\" Master's thesis, Stellenbosch University,, 2018.","[4] \tE. Beukes and J. Coetzer, \"Hand vein-based biometric authentication using two-channel similarity measure networks,\" in 2020 International SAUPEC/RobMech/PRASA Conference, University of Cape Town, Cape Town, South Africa, jan 2020, pp. 1–6.","[5] \tGarbin, X. Zhu, and O. Marques, \"Dropout vs. batch normalization: an empirical study of their impact to deep learning,\" Multimedia Tools and Applications,, vol. 79, p. 12777–12815, 2020.","[6] \tI. Goodfellow, Y. Bengio, and A. Courville, Deep Learning. MIT Press, 2016, http://www.deeplearningbook.org.","[7] \tS. Gupta and Y. Kaur, \"Review of different local and global contrast enhancement techniques for a digital image,\" International Journal of Computer Applications,, vol. 100, pp. 18–23, 2014.","[8] \tH. Hasan, H. Z. M. Shafri, and M. Habshi, \"A comparison between support vector machine (SVM) and convolu- tional neural network (CNN) models for hyperspectral image classification,\" in IOP Conference Series: Earth and Environmental Science, Volume 357, Sustainable Civil and Construction Engineering Conference,, Univer- sity Putra Malaysia, Kuala Lumpur, Malaysia, aug 2019.","[9] \tN. V. Krishnaveni, K. Sivasankari, and V. Vijayan, \"Personal authentication using hand vein,\" International Journal of Engineering Research and Technology (IJERT),, vol. 3, pp. 2333–2339, 2014. [Online]. Available: https://www.ijert.org/personal-authentication- using-hand-vein","[10] \tM. I. Obayya, M. El-Ghandour, and F. Alrowais, \"Contact- less palm vein authentication using deep learning with bayesian optimization,\" IEEE Access,, vol. 9, pp. 1940– 1957, 2021.","[11] \tL. Prechelt, \"Automatic early stopping using cross valida- tion: quantifying the criteria,\" Neural Networks, vol. 11, pp. 761–767, 1998.","[12] \tS. Processing. (2022) Credit card fraud statistics. [Online]. Available: https://shiftprocessing.com/credit-card-fraud- statistics/","[13] \tB. Sankur. (2011) Bosphorus hand vein database. Bogazici University. [Online]. Available: http://bosphorus.ee.boun.edu.tr/hand/Home.aspx","[14] \tG. Sathish, S. Narmadha, S. Saravanan, and S. U. Mah- eswari, \"Personal authentication system using hand vein biometric,\" International journal of computer technology and applications,, vol. 3, pp. 383–391, 2012.","[15] \tShorten and T. M. Khoshgoftaar, \"A survey on image data augmentation for deep learning,\" Journal of Big Data,, vol. 6, pp. 1–48, 2019."]}

Introduction. In recent years, phlebology centers have experienced an increasing inflow of patients presenting with a cosmetic issue related to bulging enlarged veins of the upper extremities. This issue especially affects people, whose jobs are associated with increased physical workloads to hands. A microfoam sclerobliteration procedure for eliminating hand veins has been performed in the phlebology center of the Medical Innovative Phlebological Center (MIPC, Moscow) since 2015. This procedure is indicated for patients with enlarged hand veins, which worsen the psycho-emotional condition of the patients. This study was aimed at developing the optimal method for conducting sclerotherapy of unaesthetic hand veins.Materials and methods. 38 patients (36 women and 2 men) have received sclerotherapy at the clinics of the Medical Innovative Phlebological Center from September 2015 to February 2020. The procedure was conducted using 1.5% to 2.5% lauromacrogol 400 microfoam. Each patient received 1 to 3 sclerotherapy sessions per hand with an interval of 3 to 7 days, the number of injections ranged from 1 to 5 per session.Results. All patients were followed over a period of 1 month to 1 year. Target veins were occluded in 100% of cases. Moderate hyperpigmentation in the projection of obliterated veins was observe in seven patients, after which it disappeared within 3 to 6 months.Conclusion. The aesthetic sclerotherapy of upper limbs is a safe and effective cosmetic procedure. After it is completed, the patients do not have any cicatrices, scars or stitches. Sclerotherapy provides demonstrates high cosmetic efficacy with minimal trauma. The procedure results in selective occlusion of targeted venous segments and makes it possible to preserve the cubital veins.

In this study, a low-cost microcomputer based dorsal hand-vein imaging system was developed by utilizing non-invasive near infrared imaging technology for the patients having hand vein visibility problems. On the microcomputer, we established a new database with high-quality dorsal-hand vein images. Pre- and post-processing were accomplished in Python language using Open-CV library on the microcomputer. Then, the specialist examined the whole obtained raw vein images to detect blood veins by marking the vein areas on a touch screen monitor, namely the reference images. The accuracy of the vein segmentation of our imaging system was assessed by comparing the reference images and the processed images on the microcomputer. We offered a novel simple pixel to pixel algorithm to test the segmented vein areas by the system. In addition, 2-D cross correlation algorithm was used for a secondary analysis. The average accuracy result of the 2-D cross correlation algorithms was 83.84% while pixel-based algorithm offered 94.47%. According to the results, we may conclude that the proposed vein imaging system do not produce suspicious unrelated vein areas which might create another clinical problem for the patients.

This paper describes in detail different hand vein recognition methods based on Wavelet-SVM, Wavelet-ANN and Image Registration. A new image segmentation and processing algorithm is proposed to efficiently locate vein regions and suitable for feature extraction (wavelet coefficients and normalized vein imagens) and classification (SVM, ANN and Image Registration). For real time recognition and high recognition rate, we proposed an integrated system which combines three above mentioned classification methods. The simulation results reveal that the proposed integrated system achieves 1% false rejection rate (FRR) and 0.02% false acceptance rate.

Large-scale data-driven DNN models have been proven to achieve great performance in various computer vision challenges, and transfer learning is proposed recently to take advantage of pre-trained DNN on a small database. Under such a framework, an innovative classification model for both identity and gender classification with hand vein information is proposed in this paper. By adopting pre-trained VGG and AlexNet model with ImageNet database and the corresponding fine-tuned ones with PolyU fingerprint and palmprint database, state-of-the-art classification results are obtained with the fine-tuned ones, which indicates that domain-specific model performs better than a generic one, and similar experimental results with faces further indicate that biometric traits share latent patterns. On the other hand, to evaluate the distribution of shared patterns, a quantized shared-index calculated as the number of correlated dictionary atoms is realized based on a sparse representation model.

Even though various dorsal hand vein pattern extraction techniques have been proposed for biometric identification, there remains considerable room for performance. This paper describes dorsal hand vein recognition using statistical and Gray Level Co-occurrence Matrix (GLCM) based features extraction techniques and artificial neural networks (ANN). For this purpose, 240 images of 80 users were obtained from Bosphorus Hand Vein Database. The images were first pre-processed by cropping region of interest (ROI), before the application of mean filtering, contrast enhancing and histogram equalizing. The ROI was then segmented by implementation of binarization method. The statistical and GLCM features were then extracted from the segmented ROI. These extracted features were sent to ANN for classification of the images. The training result shows that the proposed technique is able to recognize dorsal hand vein pattern with with considerably high accuracy of 99.32%.

In this paper the feasibility of convolutional neural networks (CNNs) for the purpose of authenticating an individual based on near infra-red (NIR) images of his/her dorsal hand vein patterns is investigated. In particular, a subset of CNNs called two-channel similarity measure networks (2CH-SMNs) is considered, since the problem at hand involves authentication (verification), instead of recognition (classification). All hand vein images are cropped and binarised before presented to the 2CH-SMN, in order to ensure that the focus of the network is on the structure of the hand veins, and not on the shape of the hand or on the grey-scale background variations surrounding the hand vein structure. A tailor-made 2CH-SMN is trained for each client enrolled into the system by considering authentic hand vein images from the client in question, as well as fraudulent images from other enrolled clients. Different images from the client in question and images from other non-training clients are employed for validation purposes, as well as for selecting optimal client-specific probability thresholds. Another distinct subset of authentic images, as well as images from imposters (non-clients) are used for testing. The results are encouraging.

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.

The region of interest (ROI) extraction is important in hand vein recognition system. The main challenges for accurate extraction of the vein region are to overcome variability in hand size, lighting conditions, orientation, appearance, noisy background, and non-uniform grey levels in foreground region. Here, we propose a new dynamic hand vein ROI extraction, preserving the whole vein area. A hand segmentation process robust to the mentioned challenges, contributing to an accurate definition of hand edge delimitations is proposed. Our approach is validated on both dorsal vein Bosphorus database and palm vein Vera database. Our proposed method accuracy is ~98% for Bosphorus database and 90% for Vera database. To illustrate the efficiency of the proposed ROI extraction, we insert it as a first block in a hand vein recognition system. Then, a comparison study at system level with recent approaches is carried on, showing an improvement of the whole system area under the curve by a rate of 12% and 2% for Bosphorus and Vera databases, respectively. The speed performances demonstrate a mean run time of 0.73 s for Bosphorus database and 1.2 s for Vera database, proving that the proposed method can be conveniently used on a real-time application.

Scale-Invariant Feature Transform (SIFT) is being investigated more and more to realize a less-constrained hand vein recognition system. Contrast enhancement (CE), compensating for deficient dynamic range aspects, is a must for SIFT based framework to improve the performance. However, evidence of negative influence on SIFT matching brought by CE is analysed by our experiments. We bring evidence that the number of extracted keypoints resulting by gradient based detectors increases greatly with different CE methods, while on the other hand the matching result of extracted invariant descriptors is negatively influenced in terms of Precision-Recall (PR) and Equal Error Rate (EER). Rigorous experiments with state-of-the-art and other CE adopted in published SIFT based hand vein recognition system demonstrate the influence. What is more, an improved SIFT model by importing the kernel of RootSIFT and Mirror Match Strategy into a unified framework is proposed to make use of the positive keypoints change and make up for the negative influence brought by CE.

An available recognition toolkit is one of the basic requirements for conducting research in finger- and hand vein recognition. Currently, there is only one comprehensive, open-source, software package available, which includes a full finger vein recognition toolkit. We present a novel, full fledged vein recognition software framework implemented in MATLAB. Our PLUS OpenVein software package contains various well-established and state-of-the-art vein enhancement, feature extraction and comparison schemes. Moreover, it contains tools to evaluate the recognition performance and provides functions to perform feature- and score-level fusion. It is fully open source and available free of charge for research and non-commercial purposes. This vein recognition framework has already been used in several papers and can be a valuable tool for other researchers working in vein recognition.

Introduction In a case of child pornography, only the dorsum of the offender's hand was clearly visible. After identification of a suspect, the question arose of whether and how it is possible to identify or exclude the suspect as perpetrator according to the morphology of the hand vein pattern. Material and methods A simple approach to use the hand vein pattern in crime suspects as a tool for identification was tested. In this study, the hand vein patterns of 30 study participants were analysed from conventional frames on videography. A standardised grid system consisting of six lines and four sectors was applied on the dorsum of the hands. Vein branchings within the sectors and line crossings of the veins were counted, leading to a total of 11 variables for each hand. Results A positive identification of each of the 30 test participants was possible for each hand when taking only the first five variables into account. A random overlapping prediction was obtained by statistically simulating hand vein patterns of different numbers of persons using this sample. Considering the hand vein frequencies in this sample, the results indicate that the chance for two persons having the same pattern is smaller than 1:1000. Conclusions It can be concluded that the introduced grid system approach can be an appropriate simple and non-costly tool for the analysis of the pattern of hand veins for identification purposes.

Hand vein recognition for personal identification has attracted considerable attention from scholars due to the uniqueness of vein pattern and nearly impossible forgery. However, recognition can be difficult to obtain when remarkable discrepancies exist in vein images caused by inconstant hand poses during imaging. This study proposes deep patch matching methodology (DPM) for hand vein recognition to solve significant differences in vein images. The proposed method calculates the similarity map of each circle patch based on vessel-enhanced filters using convolution. All similarity maps are further aggregated to calculate image similarity using max-pooling. The proposed algorithm is evaluated using public NCUT part A database. The recognition rate of the proposed algorithm is 99.71%. Experimental results reveal that the proposed algorithm obtains the highest accuracy among existing methods.

Rapid estimates of the central venous pressure (CVP) can be helpful to administer early fluid therapy or to manage cardiac preload in intensive care units, operating rooms or emergency rooms in order to start and monitor an adequate medical therapy. Invasive CVP measurements have inherent and non-negligible complication rates as well as great expenditures. Several noninvasive methods of CVP measurements, like ultrasound-guided techniques, are available, but require trained skills and special equipment which might not be at hand in all situations. Our purpose was to evaluate the feasibility and accuracy of CVP estimates assessed upon the height of hand veins collapse (HVC) using invasively measured CVP as the gold standard. The HVC was determined by slowly lifting the patient's hand while watching the dorsal hand veins to collapse. The vertical distance from the dorsal hand to a transducer air zero port was noted and converted to mmHg. The observer was blinded to the simultaneously measured CVP values, which were categorized as low (7 mmHg), normal (7-12 mmHg) and high (12 mmHg). Measurements were performed in 82 patients who had a median [IQR] age of 67 [60;74]. Median CVP was 12 [8;15] mmHg and the median absolute difference between the measured HVC and CVP was 4 [2;7] mmHg. The Spearman correlation coefficient between CVP and HVC was 0.55, 95%-CI [0.35;0.69]. Overall CVP categorization was correct in 45% of the cases. HVC had a sensitivity of 92% for a low CVP with a negative predictive value of 98%. A high HVC had a sensitivity of 29% but a high specificity of 94% for a high CVP. The overall performance of observing the hand vein collapse to estimate CVP was only moderate in the intensive care setting. However, the median difference to the CVP was low and HVC identifies a low CVP with a high sensitivity and excellent negative predictive value.

In spite of recent efforts in gathering multimodal databases containing a big number of traits, a huge amount of users and covering multiple realistic scenarios, there is still a lack of touch-less realistic samples, video recordings for some traits and the use of infrared cameras which allows, among others, to avoid lighting influence and test recently appeared biometric techniques such as hand vein recognition. For this reason, a new realistic multimodal database composed of 8,160 hand, iris and face videos has been captured. To this end, a total of 60 contributors have participated in three separated acquisition sessions in which three different cameras have been used, covering different ranges of the light spectrum: visible light and two different infrared wavelengths. To simulate real-world working conditions, the database has been recorded in an indoor environment with different lightings and backgrounds. In addition, due to the relevance of performing evaluation experiments in such a way that a reliable comparison of the results can be accomplished, an evaluation protocol is provided at the end of this paper. Moreover, performance results are provided for several biometric traits in mono- and multi- modalities that can be used as a baseline.

A new feature extraction algorithm based on polar harmonic transforms (PHT) is presented, the accomplishment of the new algorithm involves several steps. First, we design the power-controlled multispectral vein acquisition handset to establish high-quality hand vein database. Second, we propose an improved valley-shaped enhancement operator with Niblack algorithm to realize vein segmentation. Third, we design experiments to compare the capability of the several invariant moments including Hu, Zernike, PST, PCT, PCET on the feature description of the different samples and the same samples with different geometric changes. Finally we choose PCET to do the matching experiment and we also use Hu to do the experiment so as to show the superiority of the PCET. The result is that the PCET can achieve an error equation rate of 0.861% and the Hu just achieves 14.51%.

have been used from ages for authentication purpose but the key advantages of hand vein pattern that is "hand vein pattern are 100 times more unique than other physiological features" are attracting many researchers. In this paper we have proposed an efficient authentication system using hand vein pattern. The concept of information sets has been used for feature extraction so as to have control over the local information. The information set based features has already been implemented on biometrics like ear and face. Till now these features have not been implemented on hand vein pattern based verification systems hence we have explored their suitability on hand veins. All the experiments have been carried out on hand vein database developed at IIT, DELHI .We have used Hanman classifier to check the accuracy of extracted features. Then we have optimized the Hanman classifier performance by using genetic algorithm. The optimized results shows that the proposed system in 99% efficient. We have also compared Hanman classifier accuracy with accuracy obtained from k- nearest neighbor and SVM classifier. Keywordsvein, Biometrics, Information sets, Fuzzy sets, Hanman

Traditional image contrast enhancement methods originally cannot improve the quality of vein images and may also import some unknown noise resulting in low recognition rate. To overcome the abovementioned disadvantages, the paper proposes an enhancement method based on the morphological filtering theory including three main procedures. Firstly, the algorithm extract the vein Region Of Interest (ROI), and then adopting the improved White Top-Hat transform (WTH) and Black Top-Hat transform (BTH) methods to get the features of vein in detail in both white and black pattern (vein information and background information); Secondly, to construct the filtering function with the self-designed controlling operator, representing the gradient changes of the vein edges, which well reflects the importance of local detail in multi-scale pattern; Finally, traditional nonlinear gray-level transformation function is imported with modality to the parameters to realize the gray normalization. We perform rigorous experiments with the proposed method and other state-of-the-art enhancement methods on the self-built dorsal vein image databases, and the experimental results illustrate that the multiscale top-hat theory-based enhancement methods improve the contrast of hand vein images with restrictions on the possibility of enhancement on existing noise information.

Background Rocuronium-induced pain upon injection is very common in the clinical setting. Using the antecubital rather than the hand vein can avoid pain due to propofol injection. We aimed to investigate whether the use of the antecubital vein for injection would alleviate rocuronium-induced pain in a similar fashion. Methods Sixty patients (ASA classes I and II) scheduled for gynaecologic laparoscopy were randomised into two groups. Rocuronium (0.6 mg/kg) was injected either into the vein on the dorsum of the hand (group D) or a large vein in the antecubital fossa (group A). Pain was assessed and recorded using a four-point scale. Results Compared with group D, the incidence of pain and severe pain was lower in group A patients. The rate of no pain was also higher in group A patients. Conclusion The incidence and severity of rocuronium-induced injection pain were significantly alleviated via use of a large vein for rocuronium injection.

Biometrics, such as fingerprint, iris, face, hand print, hand vein, speech and gait recognition, etc., as a means of identity management has become commonplace nowadays for various applications. [...]

For the security reasons, person identification has got primary place by means of some of the physiological features. In this paper, multimodal biometry is utilised for the person identification with the help of three physiological features such as finger print, palm print and hand vein. Initially, in the pre-processing stage, the unwanted portions, noise content and the blur effects are removed from the input finger print, palm print and hand vein images. The features from these three modalities are extracted. The combined feature vector is obtained by chaff points and extracted features. After getting the combined feature vector points, the secret key points are added with the combined feature vector points to generate the fuzzy vault. Finally, in the recognition stage, test person's combined vector is compared with the fuzzy vault database. Now we can obtain the corresponding finger print, hand vein and palm print images.