Abstract:Fingerprints are popular among the biometric based systems due to ease of acquisition, uniqueness and availability. Nowadays it is used in smart phone security, digital payment and digital locker. The traditional fingerprint matching methods based on minutiae are mainly applicable for large-area fingerprint and the accuracy rate would reduce significantly when dealing with small-area fingerprint from smart phone. There are many attempts to using deep learning for small-area fingerprint recognition, and there are many successes. But training deep neural network needs a lot of datasets for training. There is no well-known dataset for small-area, so we have to make datasets ourselves. In this paper, we propose a method of data augmentation to train a small-area fingerprint recognition deep neural network with a normal fingerprint database (such as FVC2002) and verify it via tests. The experimental results showed the efficiency of our method.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is that in small - area fingerprint recognition, due to the lack of sufficient small - area fingerprint data sets, the training effect of deep neural networks is not good. Traditional fingerprint matching methods are mainly applicable to large - area fingerprints, and for small - area fingerprints collected on devices such as smart phones, their accuracy will decrease significantly. Although there have been many attempts to use deep - learning techniques for small - area fingerprint recognition and some success has been achieved, these methods usually require a large amount of training data. However, the data sets of small - area fingerprints are not abundant, and because fingerprint images are personal privacy data, it is difficult to obtain a large amount of them. In addition, cropping small - area fingerprint images from existing large - scale fingerprint databases (such as FVC2002) for training will reduce the accuracy of learning materials, thus affecting the learning effect. To solve the above problems, the paper proposes a data - augmentation - based method, using a conventional large - area fingerprint database to generate a data set suitable for training a deep neural network for small - area fingerprint recognition. The effectiveness of this method has been verified through experiments. The experimental results show that using data - augmentation techniques can effectively improve the performance of small - area fingerprint recognition. Specifically, the paper adopts multiple data - augmentation methods such as rotation, translation, brightness - contrast adjustment, and noise addition to increase the diversity of data and the generalization ability of the model.

A Method of Data Augmentation to Train a Small Area Fingerprint Recognition Deep Neural Network with a Normal Fingerprint Database

A Morphological Fingerprint Minutiae Annotation Algorithm for Deep Learning Datasets.

Data augmentation-based enhanced fingerprint recognition using deep convolutional generative adversarial network and diffusion models

Deep Belief Network Based Finger Vein Recognition Using Histograms Of Uniform Local Binary Patterns Of Curvature Gray Images

Data Augmentation for Face Recognition

Mixed Automatic Adversarial Augmentation Network for Finger-Vein Recognition

FingerNet: an Unified Deep Network for Fingerprint Minutiae Extraction.

Generative Adversarial Network (GAN) Based Data Augmentation for Palmprint Recognition

High-Resolution Mobile Fingerprint Matching Via Deep Joint Knn-Triplet Embedding

Deep Learning-inspired Automatic Minutiae Extraction from Semi-automated Annotations

A Novel 2D Contactless Fingerprint Matching Method

Robust Minutiae Extractor: Integrating Deep Networks and Fingerprint Domain Knowledge

Fingerprint Minutiae Detection Based on Multi-scale Convolution Neural Networks

Deep Palmprint Recognition with Alignment and Augmentation of Limited Training Samples

A Finger Vein Recognition Algorithm Based on Deep Learning

Deep Dense Multi-level Feature for Partial High-Resolution Fingerprint Matching

A Dataset and Benchmark for Multimodal Biometric Recognition Based on Fingerprint and Finger Vein

Fusion loss and inter-class data augmentation for deep finger vein feature learning

Synthesis of a Small Fingerprint Database through a Deep Generative Model for Indoor Localisation

A novel data augmentation method to enhance deep neural networks for detection of atrial fibrillation

DeepVein: Novel Finger Vein Verification Methods Based on Deep Convolutional Neural Networks