Related Experiment Video
Updated: Jun 27, 2026
03:31
End-To-End Deep Neural Network for Salient Object Detection in Complex Environments
Published on: December 15, 2023
Towards Fingerprint Mosaicking Artifact Detection: A Self-Supervised Deep Learning Approach
Laurenz Ruzicka1, Alexander Spenke2, Stephan Bergmann2
1Department of Digital Safety and Security, Austrian Institute of Technology, 1210 Vienna, Austria.
Sensors (Basel, Switzerland)
|June 26, 2026
Summary
This study introduces a deep learning method to detect and score fingerprint mosaicking errors, enhancing biometric system reliability. The self-supervised approach works on unlabeled data, improving image quality assessment.
Area of Science:
- Biometrics
- Computer Vision
- Machine Learning
Background:
- Fingerprint mosaicking combines multiple impressions but is prone to errors degrading image quality.
- Accurate detection of these hard mosaicking artifacts is crucial for reliable biometric systems.
Purpose of the Study:
- To develop a deep learning-based method for detecting and scoring hard mosaicking artifacts in fingerprint images.
- To enable automated evaluation of fingerprint image quality at scale.
Main Methods:
- A self-supervised learning framework was used to train a segmentation model on large-scale unlabeled fingerprint data.
- The model was evaluated across various fingerprint modalities (contactless, rolled, pressed) and data sources.
- A novel mosaicking artifact score was introduced to quantify error severity.
Main Results:
- The proposed model achieved high segmentation performance in identifying mosaicking errors.
- The method demonstrated robustness across different fingerprint types and data sources.
- The mosaicking artifact score enables scalable, automated quality assessment.
Conclusions:
- The deep learning approach effectively addresses the challenge of reference-free hard artifact detection in fingerprint mosaicking.
- This work contributes to improving the accuracy and reliability of fingerprint-based biometric systems.