Thansk for sharing the paper and code. It is a good work.
I notice your work is closely related to MATE (Masked Autoencoders are Online 3D Test-Time Learners, ICCV'23) and BFTT3D (Backpropagation-free Network for 3D Test-time Adaptation, CVPR'24), but your method did not explicitly compare with them.
Thansk for sharing the paper and code. It is a good work.
I notice your work is closely related to MATE (Masked Autoencoders are Online 3D Test-Time Learners, ICCV'23) and BFTT3D (Backpropagation-free Network for 3D Test-time Adaptation, CVPR'24), but your method did not explicitly compare with them.
Can you provide some insights? Thanks.