Reproduce “Deep Heterogeneous Feature Fusion for Template-Based Face Recognition” 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

Navaneeth Bodla, Jingxiao Zheng, Hongyu Xu, Jun-Cheng Chen, Carlos Castillo, Rama Chellappa

The authors develop a face-recognition system that fuses features from two pretrained networks for verification of templates containing multiple images or video frames. Each network’s features are averaged within a template before learned nonlinear fusion. On Janus CS3 mixed-media verification, the softmax-trained system reaches TPR 0.6453 at FPR 0.0001, versus 0.6051 for score-level averaging, under this protocol.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.