LatentPrintFormer: A Hybrid CNN-Transformer with Spatial Attention for Latent Fingerprint identification

2025-11-11

AI Summary

Latent fingerprint identification remains challenging due to low-quality images, strong background noise, and incomplete ridge patterns. To address these issues, this paper proposes a hybrid CNN-Transformer architecture featuring dual backbones—EfficientNet-B0 and Swin-Tiny—integrated with a spatial attention module to enhance discriminative representation of high-quality ridge regions. Local and global features are jointly optimized via complementary fusion and contrastive projection, enabling closed-set matching through cosine similarity. Extensive experiments on two public latent fingerprint datasets demonstrate that the proposed method achieves significantly higher Rank-10 identification accuracy than three state-of-the-art approaches. The results validate its effectiveness in modeling complex latent fingerprint characteristics and performing robust, discriminative matching under challenging conditions.

Technology Category

• Computer Vision: Other Foundations of Computer Vision
• Intelligent Robots: Multimodal Perception & Sensor Fusion
• Machine Learning: Feature Construction/Reformulation

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Security and Privacy: Data transparency and provenance
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Latent fingerprint identification remains a challenging task due to low image quality, background noise, and partial impressions. In this work, we propose a novel identification approach called LatentPrintFormer. The proposed model integrates a CNN backbone (EfficientNet-B0) and a Transformer backbone (Swin Tiny) to extract both local and global features from latent fingerprints. A spatial attention module is employed to emphasize high-quality ridge regions while suppressing background noise. The extracted features are fused and projected into a unified 512-dimensional embedding, and matching is performed using cosine similarity in a closed-set identification setting. Extensive experiments on two publicly available datasets demonstrate that LatentPrintFormer consistently outperforms three state-of-the-art latent fingerprint recognition techniques, achieving higher identification rates across Rank-10.

Problem

• Improving latent fingerprint identification accuracy
• Handling low-quality images with noise suppression
• Integrating local and global feature extraction

Innovation

• Hybrid CNN-Transformer architecture for feature extraction
• Spatial attention module to enhance ridge regions
• Unified embedding with cosine similarity matching