Fingerprint Saliency Dataset
收藏资源简介:
该数据集名为Fingerprint Saliency Dataset,由圣母大学计算机科学与工程学院的研究团队创建。数据集包含800个人工标注的指纹感知重要图,这些图与文本描述一起,支持人类检测伪造指纹的决策。此外,数据集还包括算法生成的伪显著图,包括基于细节、图像质量和自动编码器生成的显著图。该数据集旨在用于研究显著性引导训练在指纹活体检测中的应用,以提高模型在有限和大量数据环境中的泛化能力和分类精度。
This dataset, named Fingerprint Saliency Dataset, was created by a research team from the Department of Computer Science and Engineering, University of Notre Dame. The dataset contains 800 manually annotated fingerprint saliency maps, which, together with textual descriptions, support human decision-making for fake fingerprint detection. Additionally, the dataset also includes algorithmically generated pseudo saliency maps based on fingerprint minutiae, image quality, and autoencoders. This dataset is intended for research on the application of saliency-guided training in fingerprint liveness detection, aiming to improve the model's generalization ability and classification accuracy in both limited and large-scale data scenarios.

- 1Saliency-Guided Training for Fingerprint Presentation Attack Detection圣母大学计算机科学与工程学院 · 2025年



