Submitted by Weiwei Sun.

Submission data

Full nameNeural Bilateral Filtering for Top-down Instance Segmentation on Point Clouds
DescriptionWe introduce a method for instance proposal generation for 3D point clouds. Existing techniques typically directly regress proposals in a single feed-forward step, leading to inaccurate estimation. We show that this serves as a critical bottleneck, and propose a method based on iterative bilateral filtering with learned kernels. Following the spirit of bilateral filtering, we consider both the deep feature embeddings of each point, as well as their locations in the 3D space. We show via synthetic experiments that our method brings drastic improvements when generating instance proposals for a given point of interest. We further validate our method on the challenging ScanNet benchmark, achieving the best instance segmentation performance amongst the sub-category of top-down methods.
Publication titleNeuralBF: Neural Bilateral Filtering for Top-down Instance Segmentation on Point Clouds
Publication authorsWeiwei Sun, Daniel Rebain, Renjie Liao, Vladimir Tankovich, Soroosh Yazdani, Kwang Moo Yi, Andrea Tagliasacchi
Publication venueWACV 2023
Publication URLhttps://neuralbf.github.io/
Input Data TypesUses Color,Uses Geometry        Uses 3D
Programming language(s)Python/Pytorch
HardwareNVIDIA 3090
Websitehttps://neuralbf.github.io/
Source code or download URLhttps://neuralbf.github.io/
Submission creation date9 Jul, 2022
Last edited10 Jan, 2023

3D semantic instance results



Infoavg ap 25%bathtubbedbookshelfcabinetchaircountercurtaindeskdoorotherfurniturepicturerefrigeratorshower curtainsinksofatabletoiletwindow
0.7181.0000.9450.9010.7540.8170.4600.7000.7720.6880.5680.0000.5000.9810.6060.8720.7401.0000.614