Search papers, labs, and topics across Lattice.
4
5
4
27
OpenVeinNet achieves unprecedented cross-dataset generalization in finger vein verification by combining advanced feature extraction with graph learning techniques.
Face recognition systems, beware: DCMorph's dual-stream diffusion morphs achieve unprecedented attack success rates while remaining stealthy to existing detection methods.
A unified benchmark reveals the trade-offs between pixel-wise accuracy and perceptual realism in state-of-the-art image super-resolution techniques.
Speaker verification systems are shockingly vulnerable: a new signal-level voice morphing attack achieves near-perfect success rates (G-MAP up to 99.74%) against both deep learning and commercial systems.
