Reproduce “Lightweight graph transformers for clutter and target classification in automotive radar” 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
Adrian Gheorghiu, Tunc Alkanat, Ashish Pandharipande
The paper presents TGTNet, a compact point-cloud transformer that classifies radar detections as clutter, stationary, or moving targets. It combines local attention with geometry-aware clustering and also tests multi-frame inputs. On RadarScenes, dimension-32 single-frame TGTNet reaches 94.46 macro-mean F1, versus 95.45 for GRT, with 80.25 KB size and 6.83 ms latency; this reduced-feature GPU comparison does not establish automotive deployment.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.