search

hailanyi / TED

Transformation-Equivariant 3D Object Detection for Autonomous Driving
https://arxiv.org/abs/2211.11962
Apache License 2.0
139 stars 32 forks source link

Transformation scale #28

Open minju-Kang opened 1 year ago

minju-Kang commented 1 year ago

Hi, thank you for great work.

Is there any reason you fixed the scale of transformation for each transformation channel? Is there any ablation study with randomly transformed channel?