FV-GAN: Finger Vein Representation Using Generative Adversarial Networks

Yang, Wenming; Hui, Changqing; Chen, Zhiquan; Xue, Jing-Hao; Liao, Qingmin

doi:10.1109/tifs.2019.2902819

Cited by 114 publications

(56 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In work [39], a feature extraction and recovery network, which outperformed manual feature extraction in terms of verification errors, was proposed. As an extension of this work, [40] was engaged with an novel method to extract the depth feature of vein based on both short-term and long-term memory recurrent neural network, while Yang et al [41] introduced a finger vein segmentation model in light of the generative adversarial networks, which stands astonishing robustness to outliers and vessel breaks. Referring to [42]- [44], semantic segmentation convolutional neural network was optimized to directly abstract the actual finger-vein patterns from NIR finger images while Jalilian and Uhl [43] investigated the effects of fusion and combination training with different labels on finger vein recognition.…”

Section: Related Work and Motivationmentioning

confidence: 99%

Finger Vein Verification Algorithm Based on Fully Convolutional Neural Network and Conditional Random Field

et al. 2020

View full text Add to dashboard Cite

Owing to the complexity of finger vein patterns in shape and spatial dependence, the existing methods suffer from an inability to obtain accurate and stable finger vein features. This paper, so as to compensate this defect, proposes an end-to-end model to extract vein textures through integrating the fully convolutional neural network (FCN) with conditional random field (CRF). Firstly, to reduce missing pixels during ROI extraction, the method of sliding window summation is employed to filter and adjusted with self-built tools. In addition, the traditional baselines are endowed with different weights to automatically assign labels. Secondly, the deformable convolution network, through replacing the plain counterparts in the standard U-Net mode, can capture the complex venous structural features by adaptively adjusting the receptive fields according to veins' scales and shapes. Moreover, the above features can be further mined and accumulated by combining the recurrent neural network (RNN) and the residual network (ResNet). With the steps mentioned above, the fully convolutional neural network is constructed. Finally, the CRF with Gaussian pairwise potential conducts mean-field approximate inference as the RNN, and then is embedded as a part of the FCN, so that the model can fully integrate CRF with FCNs, which provides the possibility to involve the usual back-propagation algorithm in training the whole deep network end-to-end. The proposed models in this paper were tested on three public finger vein datasets SDUMLA, MMCBNU and HKPU with experimental results to certify their superior performance on finger-vein verification tasks compared with other equivalent models including U-Net.

show abstract

Section: Related Work and Motivationmentioning

confidence: 99%

Finger Vein Verification Algorithm Based on Fully Convolutional Neural Network and Conditional Random Field

et al. 2020

View full text Add to dashboard Cite

show abstract

“…Recently, some more powerful but complex network models, such as Siamese Network [ 46 ], GaborPCA Network [ 47 , 48 ], Convolutional Autoencoder [ 49 ], Capsule Network [ 50 ], DenseNet [ 51 , 52 ], Fully Convolutional Network (FCN) [ 53 , 54 ], Generative Adversarial Network (GAN) [ 55 , 56 , 57 ], and Long Short-term Memory (LSTM) Network [ 58 ], etc., also emerged in the field of FV image recognition. Especially for the GAN, which can not only achieve more robust vein patterns from low-quality FV images, but also generate a variety of synthetic FV samples.…”

Section: Introductionmentioning

confidence: 99%

A Novel Finger Vein Recognition Method Based on Aggregation of Radon-Like Features

Yao

Song

et al. 2021

Sensors

View full text Add to dashboard Cite

Finger vein (FV) biometrics is one of the most promising individual recognition traits, which has the capabilities of uniqueness, anti-forgery, and bio-assay, etc. However, due to the restricts of imaging environments, the acquired FV images are easily degraded to low-contrast, blur, as well as serious noise disturbance. Therefore, how to extract more efficient and robust features from these low-quality FV images, remains to be addressed. In this paper, a novel feature extraction method of FV images is presented, which combines curvature and radon-like features (RLF). First, an enhanced vein pattern image is obtained by calculating the mean curvature of each pixel in the original FV image. Then, a specific implementation of RLF is developed and performed on the previously obtained vein pattern image, which can effectively aggregate the dispersed spatial information around the vein structures, thus highlight vein patterns and suppress spurious non-boundary responses and noises. Finally, a smoother vein structure image is obtained for subsequent matching and verification. Compared with the existing curvature-based recognition methods, the proposed method can not only preserve the inherent vein patterns, but also eliminate most of the pseudo vein information, so as to restore more smoothing and genuine vein structure information. In order to assess the performance of our proposed RLF-based method, we conducted comprehensive experiments on three public FV databases and a self-built FV database (which contains 37,080 samples that derived from 1030 individuals). The experimental results denoted that RLF-based feature extraction method can obtain more complete and continuous vein patterns, as well as better recognition accuracy.

show abstract

“…In this article, the proposed method employed generative adversarial networks (GANs) to develop an audio imitation training system. GANs, 5–8 first proposed by Goodfellow et al, 9 employ two deep neural networks for adversarial training. In this unique training method, the original latent space is mapped onto the actual data distribution.…”

Section: Introductionmentioning

confidence: 99%

DNAE-GAN: Noise-free acoustic signal generator by integrating autoencoder and generative adversarial network

Kuo

Lin

2020

International Journal of Distributed Sensor Networks

View full text Add to dashboard Cite

Linear predictive coding is an extremely effective voice generation method that operates through simple process. However, linear predictive coding–generated voices have limited variations and exhibit excessive noise. To resolve these problems, this article proposes an artificial intelligence model that combines a denoise autoencoder with generative adversarial networks. This model generates voices with similar semantics through the random input from the latent space of generator. The experimental results indicate that voices generated exclusively by generative adversarial networks exhibit excessive noise. To solve this problem, a denoise autoencoder was connected to the generator for denoising. The experimental results prove the feasibility of the proposed voice generation method. In the future, this method can be applied in robots and voice generation applications to increase the humanistic language expression ability of robots and enable robots to demonstrate more humanistic and natural speaking performance.

show abstract

FV-GAN: Finger Vein Representation Using Generative Adversarial Networks

Cited by 114 publications

References 31 publications

Finger Vein Verification Algorithm Based on Fully Convolutional Neural Network and Conditional Random Field

Finger Vein Verification Algorithm Based on Fully Convolutional Neural Network and Conditional Random Field

A Novel Finger Vein Recognition Method Based on Aggregation of Radon-Like Features

DNAE-GAN: Noise-free acoustic signal generator by integrating autoencoder and generative adversarial network

Contact Info

Product

Resources

About