Diff-Palm: Realistic Palmprint Generation with Polynomial Creases and Intra-Class Variation Controllable Diffusion Models

2025-03-24

AI Summary

Palmprint recognition is severely hindered by the scarcity of large-scale, publicly available real-world datasets; existing Bézier-curve-based synthesis methods fail to simultaneously preserve intra-class diversity and identity consistency, leading to poor model generalization. Method: We propose the first realistic palmprint generation framework based on polynomial modeling: (i) a novel polynomial-based wrinkle representation that accurately captures palmprint geometric structure; (ii) a wrinkle-conditioned diffusion model coupled with a K-step noise-sharing sampling mechanism to enable controllable, identity-consistent intra-class variation. Contribution/Results: A recognition model trained solely on synthetic data achieves, for the first time, performance surpassing that trained on real data—and further improves monotonically as the number of generated identities increases—effectively alleviating the data scarcity bottleneck.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Learning with Manifolds
• Natural Language Processing: Generation

Application Category

• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Responsible Web: Data and user privacy-enhancing technologies for the Web
• Web Mining and Content Analysis: Web data generation and simulation

Abstract

Palmprint recognition is significantly limited by the lack of large-scale publicly available datasets. Previous methods have adopted B'ezier curves to simulate the palm creases, which then serve as input for conditional GANs to generate realistic palmprints. However, without employing real data fine-tuning, the performance of the recognition model trained on these synthetic datasets would drastically decline, indicating a large gap between generated and real palmprints. This is primarily due to the utilization of an inaccurate palm crease representation and challenges in balancing intra-class variation with identity consistency. To address this, we introduce a polynomial-based palm crease representation that provides a new palm crease generation mechanism more closely aligned with the real distribution. We also propose the palm creases conditioned diffusion model with a novel intra-class variation control method. By applying our proposed $K$-step noise-sharing sampling, we are able to synthesize palmprint datasets with large intra-class variation and high identity consistency. Experimental results show that, for the first time, recognition models trained solely on our synthetic datasets, without any fine-tuning, outperform those trained on real datasets. Furthermore, our approach achieves superior recognition performance as the number of generated identities increases.

Problem

• Lack of large-scale public palmprint datasets limits recognition accuracy
• Existing synthetic palmprints fail to match real data quality
• Balancing intra-class variation and identity consistency is challenging

Innovation

• Polynomial-based palm crease representation
• Diffusion model with intra-class control
• K-step noise-sharing sampling method