Progressive Learning of a Diffusion-based Inpainting Model for Separating Overlapped Fingerprints

2026-08-04

AI Summary

This work addresses the challenging problem of separating overlapping fingerprints, a task for which existing approaches either rely heavily on strong domain-specific knowledge or neglect intrinsic fingerprint structural characteristics. The authors formulate fingerprint separation as an image inpainting task and propose an overlap-aware diffusion-based inpainting framework. Built upon a pretrained Stable Diffusion model, their method guides the diffusion process through multi-channel conditioning, integrates fingerprint orientation field priors, and employs a staged progressive fine-tuning strategy to achieve high-fidelity reconstruction of individual fingerprint components. This study presents the first integration of progressive learning with a multi-channel conditional diffusion model for fingerprint separation, significantly improving matching success rates between reconstructed and ground-truth fingerprints on two public datasets. The approach effectively combines domain knowledge with data-driven learning, demonstrating notable practical potential.

Abstract

Overlapped friction ridge patterns are a recurring problem in latent fingerprints recovered from crime scenes and in live-scan scenarios where residual fingerprints on the sensor may corrupt subsequent acquisitions. Existing approaches for separating overlapped fingerprints either rely on rule-based orientation field completion that requires strong domain knowledge or train end-to-end deep neural networks that do not account for domain-specific considerations. This work introduces a diffusion-based pipeline for separating component fingerprints from an image containing overlapping friction ridge patterns. We formulate the separation problem as an inpainting task and progressively learn a diffusion model for this task in multiple stages. Starting from a pre-trained Stable Diffusion model, we progressively incorporate a fingerprint prior, add the ability to complete partial fingerprints, and finally propose \textbf{overlap-aware inpainting} that reconstructs each component print using a diffusion inpainting model based on multi-channel conditioning. Experiments on two public datasets demonstrate that component fingerprints reconstructed using the proposed diffusion-based inpainting method can match with their mated counterparts with very high probability.

Problem

• overlapped fingerprints
• fingerprint separation
• latent fingerprints
• friction ridge patterns
• inpainting

Innovation

• diffusion-based inpainting
• overlap-aware inpainting
• progressive learning
• fingerprint separation
• multi-channel conditioning