search

junjie18 / CMT

[ICCV 2023] Cross Modal Transformer: Towards Fast and Robust 3D Object Detection
Other
308 stars 34 forks source link

Robustness Study #102

Open TillBeemelmanns opened 3 months ago

TillBeemelmanns commented 3 months ago

Hi,

thanks for this interesting work! I did a small robustness study on multi-modal object detectors and found out, that CMT currently outperforms all other models on our benchmark. Maybe you are interested in this.

Code: https://github.com/ika-rwth-aachen/MultiCorrupt Preprint: https://arxiv.org/abs/2402.11677

Our results will be published in the IEEE Intelligent Vehicles Symposium, June 2024.

Thanks!

Bests, Till