Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
2
0
5
16
You don't need deep learning to achieve 93% accuracy in detecting deepfake audio – a well-tuned SVM with interpretable acoustic features gets you there.
Smartphone iris recognition in visible light can now achieve surprisingly high accuracy (97.9% TAR at 0.01 FAR) thanks to a new dataset, acquisition protocol, and lightweight VIS-adapted models.