Contactless Palm Vein Authentication Using Deep Learning With Bayesian Optimization

Obayya, Marwa; El-Ghandour, Mohammed; Alrowais, Fadwa

doi:10.1109/access.2020.3045424

Cited by 18 publications

(10 citation statements)

References 51 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The originality and novelty of the proposed approach stems from the new architecture assigned to recognizing people according to their palmvein images. The approach starts after the segmentation of the ROI from the palmvein image at the input according to the same segmentation steps in our previous work [64]. Figure 3 demonstrates the main steps of our proposed approach which includes feature extraction and selection, dimensionality reduction using PCA, and recognition using a deep neural network with Bayesian optimization.…”

Section: Methodsmentioning

confidence: 99%

Palmvein Recognition Using Block-Based WLD Histogram of Gabor Feature Maps and Deep Neural Network With Bayesian Optimization

et al. 2021

Self Cite

View full text Add to dashboard Cite

The demand on devices and systems empowered by biometric verification and identification mechanisms has been increasing in recent years as they have become a significant part of our lives. Palm vein biometric is an emerging technology that has drawn considerable attention from researchers and scientists over the last decade. In the present work, a novel feature extraction methodology named GPWLD combing the Gabor features with positional Weber's local descriptor (PWLD) features is proposed. WLD is a highly representative micro-pattern descriptor that performs well against noise and illumination changes in images. However, it lacks the ability to capture the vein pattern features at different orientations. Moreover, its descriptor packs the local information content into a single histogram that does not take the spatial locality of micro-patterns into consideration. To solve these two issues, firstly, the palm vein image is passed through Gabor filter with different orientations to capture the salient rotational features found in the output feature maps. Secondly, the spatiality is achieved by uniformly dividing each feature map into several blocks. Next, Weber's law feature descriptor (WLD) histogram is computed for each block in every feature map. Finally, these histograms are concatenated to compose the final feature vector. Due to the high dimensionality of the final feature vector, Principal component analysis (PCA) algorithm is utilized for feature size reduction. In the classification stage, a deep neural network (DNN) comprising an optimized stacked autoencoder (SAE) with Bayesian optimization and a softmax layer is used. Optimization of the SAE is carried out by using Bayesian optimization to find the optimal SAE structure and the options of the training algorithm. The Experimental results verify the discriminative power of the extracted features and the accuracy of the proposed DNN. For both Identification and verification experiments, the proposed method attains higher identification rate and lower EER than state-of-the-art methods.INDEX TERMS WLD descriptor, Gabor filter, deep neural network, stacked autoencoder, Bayesian optimization.

show abstract

Section: Methodsmentioning

confidence: 99%

Palmvein Recognition Using Block-Based WLD Histogram of Gabor Feature Maps and Deep Neural Network With Bayesian Optimization

et al. 2021

Self Cite

View full text Add to dashboard Cite

show abstract

“…Table 2 also includes two works from 2018 and 2019, [48] and [49], where ResNet50/ResNet101 and DenseNet161 [50] were implemented on the public SDUMLA finger vein dataset. Other studies, such as [51] on the CASIA [52] image set, proposed their owndesigned CNNs.…”

Section: ) Datasetsmentioning

confidence: 99%

Vision Transformers for Vein Biometric Recognition

Garcia-Martin

Sánchez-Reillo

2023

IEEE Access

View full text Add to dashboard Cite

In October 2020, Google researchers present a promising Deep Learning architecture paradigm for Computer Vision that outperforms the already standard Convolutional Neural Networks (CNNs) on multiple image recognition state-of-the-art datasets: Vision Transformers (ViTs). Based on the self-attention concept inherited from Natural Language Processing (NLP), this new structure surpasses the CNN image classification task on ImageNet, CIFAR-100, and VTAB, among others, when it is finetuned (Transfer Leaning) after a previous pre-training on larger datasets. In this work, we confirm this theory and move one step further over the CNN structures applied for Vascular Biometric Recognition (VBR): to the best of our knowledge, we introduce for the first time multiple pure pre-trained and finetuned Vision Transformers in this evolving biometric modality to address the challenge of the limited number of samples in VBR datasets. For this purpose, the ViTs have been trained to extract unique image features on the ImageNet-1k and ImageNet-21k and then fine-tuned for the four main existing VBR variants, i.e., finger, palm, hand dorsal, and wrist vein areas. Fourteen existing vascular datasets have been used to perform the vein identification task in the four previously mentioned modalities, based on the True-Positive Identification Rate (TPIR) and 75-25% train-test sets obtaining the following results: HKPU (99.52%), and FV-USM (99.1%); Vera (99.39%), and CASIA (96.00%); Bosphorus (99.86%); PUT-wrist (99.67%), and UC3M-CV1+CV2 (99.67%). Furthermore, we introduce UC3M-CV3: a hygienic contactless wrist database collected on smartphones and consisting of 5400 images from 100 different subjects. The promising results show the Vision Transformer's versatility in VBR under Transfer Learning and reinforce this new Neural Network architecture paradigm.INDEX TERMS Vision Transformers, vein biometric recognition, deep learning, convolutional neural networks, finger veins, transfer learning, machine learning, artificial intelligence, biometrics on mobile devices, contactless wrist vascular database, hand palm identification.

show abstract

“…In the current work, the CLAHE algorithm is also tested and compared with raw images. Obayya et al [18] present a contactless palm solution using the CASIA Multispectral Palmprint Image database [19] for CNN training and testing. A CNN architecture is designed and trained by the authors using a Bayesian optimization to find the optimal network structure and its parameters.…”

Section: ) Acquisition Storage and Processing Hardwarementioning

confidence: 99%

Deep Learning for Vein Biometric Recognition on a Smartphone

Garcia-Martin

Sánchez-Reillo

2021

IEEE Access

View full text Add to dashboard Cite

The ongoing COVID-19 pandemic has pointed out, even more, the important need for hygiene contactless biometric recognition systems. Vein-based devices are great non-contact options although they have not been entirely well-integrated in daily life. In this work, in an attempt to contribute to the research and development of these devices, a contactless wrist vein recognition system with a reallife application is revealed. A Transfer Learning (TL) method, based on different Deep Convolutional Neural Networks architectures, for Vascular Biometric Recognition (VBR), has been designed and tested, for the first time in a research approach, on a smartphone. TL is a Deep Learning (DL) technique that could be divided into networks as feature extractor, i.e., using a pre-trained (different large-scale dataset) Convolutional Neural Network (CNN) to obtain unique features that then, are classified with a traditional Machine Learning algorithm, and fine-tuning, i.e., training a CNN that has been initialized with weights of a pre-trained (different large-scale dataset) CNN. In this study, a feature extractor base method has been employed. Several architecture networks have been tested on different wrist vein datasets: UC3M-CV1, UC3M-CV2, and PUT. The DL model has been integrated on the Xiaomi© Pocophone F1 and the Xiaomi© Mi 8 smartphones obtaining high biometric performance, up to 98 % of accuracy and less than 0.4 % of EER with a 50-50 % train-test on UC3M-CV2, and fast identification/verification time, less than 300 milliseconds. The results infer, high DL performance and integration reachable in VBR without direct user-device contact, for real-life applications nowadays.

show abstract

Contactless Palm Vein Authentication Using Deep Learning With Bayesian Optimization

Cited by 18 publications

References 51 publications

Palmvein Recognition Using Block-Based WLD Histogram of Gabor Feature Maps and Deep Neural Network With Bayesian Optimization

Palmvein Recognition Using Block-Based WLD Histogram of Gabor Feature Maps and Deep Neural Network With Bayesian Optimization

Vision Transformers for Vein Biometric Recognition

Deep Learning for Vein Biometric Recognition on a Smartphone

Contact Info

Product

Resources

About