Universal Fingerprint Generation: Controllable Diffusion Model with Multimodal Conditions

TL;DR

GenPrint introduces a controllable multimodal diffusion framework for synthetic fingerprint generation that preserves finger identity while allowing explicit control over appearance factors such as class, acquisition, sensor, and quality. By incorporating text prompts and style embeddings, plus a ridge-based identity-preservation network, GenPrint supports zero-shot style generation for unseen sensors without extra fine-tuning. Extensive experiments show GenPrint achieves realism comparable to real data, improves recognition models trained on synthetic data, and enables robust evaluation while maintaining low identity leakage. The approach yields a scalable source of diverse synthetic fingerprints, with potential applicability to other biometric modalities.

Abstract

The utilization of synthetic data for fingerprint recognition has garnered increased attention due to its potential to alleviate privacy concerns surrounding sensitive biometric data. However, current methods for generating fingerprints have limitations in creating impressions of the same finger with useful intra-class variations. To tackle this challenge, we present GenPrint, a framework to produce fingerprint images of various types while maintaining identity and offering humanly understandable control over different appearance factors such as fingerprint class, acquisition type, sensor device, and quality level. Unlike previous fingerprint generation approaches, GenPrint is not confined to replicating style characteristics from the training dataset alone: it enables the generation of novel styles from unseen devices without requiring additional fine-tuning. To accomplish these objectives, we developed GenPrint using latent diffusion models with multimodal conditions (text and image) for consistent generation of style and identity. Our experiments leverage a variety of publicly available datasets for training and evaluation. Results demonstrate the benefits of GenPrint in terms of identity preservation, explainable control, and universality of generated images. Importantly, the GenPrint-generated images yield comparable or even superior accuracy to models trained solely on real data and further enhances performance when augmenting the diversity of existing real fingerprint datasets.

Figures (13)

• Figure 1: Synthetic fingerprint images generated by various baseline methods and the proposed GenPrint. The four images in each panel are impressions of the same finger to show case the intra-class variance of each method.
• Figure 2: Architecture of GenPrint.
• Figure 3: AFR-Net similarity score distributions for NIST SD302 real dataset and similar GenPrint dataset.
• Figure 4: Example GenPrint images of different fingerprint classes and corresponding classification accuracy of Verifinger v12.4 SDK.
• Figure 5: T-SNE plots to show (a) separation of GenPrint-generated images from different acquisition devices, (b) similarity of GenPrint images and corresponding real images of the same acquisition device, (c) similarity of zero-shot generated images to corresponding real images of novel acquisition devices which were not included in the training set of GenPrint.