Any3DIS: Class-Agnostic 3D Instance Segmentation by 2D Mask Tracking

TL;DR

Any3DIS introduces a novel 3D-aware 2D mask tracking approach that leverages robust priors and dynamic view selection to improve class-agnostic 3D instance segmentation accuracy and reduce redundant proposals.

Contribution

The paper presents a new 3D Mask Optimization module and a 3D-Aware 2D Mask Tracking method that enhance segmentation consistency and proposal quality over prior unsupervised merging techniques.

Findings

Improves 3D instance segmentation accuracy on ScanNet datasets.

Reduces redundant and over-segmented proposals.

Enhances performance in open-vocabulary segmentation tasks.

Abstract

Existing 3D instance segmentation methods frequently encounter issues with over-segmentation, leading to redundant and inaccurate 3D proposals that complicate downstream tasks. This challenge arises from their unsupervised merging approach, where dense 2D instance masks are lifted across frames into point clouds to form 3D candidate proposals without direct supervision. These candidates are then hierarchically merged based on heuristic criteria, often resulting in numerous redundant segments that fail to combine into precise 3D proposals. To overcome these limitations, we propose a 3D-Aware 2D Mask Tracking module that uses robust 3D priors from a 2D mask segmentation and tracking foundation model (SAM-2) to ensure consistent object masks across video frames. Rather than merging all visible superpoints across views to create a 3D mask, our 3D Mask Optimization module leverages a dynamic programming algorithm to select an optimal set of views, refining the superpoints to produce a final 3D proposal for each object. Our approach achieves comprehensive object coverage within the scene while reducing unnecessary proposals, which could otherwise impair downstream applications. Evaluations on ScanNet200 and ScanNet++ confirm the effectiveness of our method, with improvements across Class-Agnostic, Open-Vocabulary, and Open-Ended 3D Instance Segmentation tasks.