NeuralBF: Neural Bilateral Filtering for Top-down Instance Segmentation on Point Clouds

TL;DR

NeuralBF introduces an iterative bilateral filtering approach with learned kernels for improved 3D point cloud instance proposal generation, significantly enhancing accuracy over direct regression methods.

Contribution

The paper proposes a novel iterative bilateral filtering method with learned kernels for 3D point cloud instance proposals, outperforming existing direct regression techniques.

Findings

Drastic improvements in synthetic experiments

Achieves best performance among top-down methods on ScanNet

Demonstrates effectiveness of bilateral filtering in 3D segmentation

Abstract

We 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.