Reproduce “Foundation and Multimodal Large Language Models for Face Presentation and Morph Attack Detection” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Hatef Otroshi Shahreza, Asif Hussain Khan, Peter Lorenz, Alain Komaty, Sébastien Marcel
The authors test pretrained vision models for presentation-attack and morph-attack detection across datasets, comparing prompting, frozen features, and adaptation. Their strongest approach adds low-rank adapters to a vision encoder and trains a classifier. Cross-dataset ACER is 19.8% versus 22.9% for the PAD specialist and 4.9% versus 12.9% for MAD, but unseen attack types and image domains remain untested.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.