GaitEditor ermöglicht das Bearbeiten von Fußgängerattributen für Gangerkennung und Anonymisierung.
Utilizing 3D spatio-spatial temporal convolution, FH-SSTNet offers superior performance in forehead-based user verification.
Integrating handcrafted minutia features with deep neural network embeddings significantly improves latent fingerprint recognition results.
Novel approaches using GANs and diffusion models for high-quality fingerprint image synthesis.
Efficiently estimate identity, gender, and age from hand images for criminal investigations.
Deep learning-based model for non-linear iris texture deformation preserves identity and enhances recognition performance.
Mixing IrisCodes using bitwise operators can lead to effective dictionary attacks, generating alpha-wolves and alpha-mammals.
Exploring the feasibility of kinship verification using remote photoplethysmography signals.
Using synthetic images in morphing attack detection can improve performance but is not sufficient for real-world scenarios.
The author presents a deep learning-based model for non-linear iris texture deformation, preserving identity across varying pupil sizes, enhancing biometric performance.