The field of biometric identification and image processing is rapidly evolving, with a focus on improving accuracy, robustness, and scalability. Recent developments have centered around innovative approaches to feature extraction, matching, and learning frameworks. Notable advancements include the use of saliency-guided training, matchability-based reweighting, and robust duality learning to mitigate the effects of noisy pseudo-labels and improve model generalization capabilities. Additionally, novel methods for finger pose estimation, tooth segmentation, and image alignment have demonstrated significant improvements over existing state-of-the-art approaches. These advancements have the potential to enhance the performance and reliability of various biometric identification systems and image processing applications.

Noteworthy papers include: Saliency-Guided Training for Fingerprint Presentation Attack Detection, which demonstrates the effectiveness of saliency-guided training for fingerprint PAD. Focus What Matters: Matchability-Based Reweighting for Local Feature Matching, which proposes a novel attention reweighting mechanism that simultaneously incorporates a learnable bias term and applies a matchability-informed rescaling to the input value features. Finger Pose Estimation for Under-screen Fingerprint Sensor, which presents a novel dual-modal input-based network for under-screen fingerprint pose estimation. RaIL: Region-Aware Instructive Learning for Semi-Supervised Tooth Segmentation in CBCT, which promotes intergroup knowledge transfer and collaborative region-aware instruction while reducing overfitting to the characteristics of any single model.