Open3DIS: Open-Vocabulary 3D Instance Segmentation with 2D Mask Guidance

TL;DR

Open3DIS introduces a method that leverages 2D mask aggregation across frames to improve open-vocabulary 3D instance segmentation, especially for small and ambiguous objects, achieving state-of-the-art results.

Contribution

The paper presents a novel module that combines 2D mask aggregation with 3D proposals to enhance open-vocabulary 3D instance segmentation performance.

Findings

Significant performance improvements on ScanNet200, S3DIS, and Replica datasets.

Effective segmentation of small-scale and geometrically ambiguous objects.

Outperforms existing state-of-the-art methods in open-vocabulary 3D segmentation.

Abstract

We introduce Open3DIS, a novel solution designed to tackle the problem of Open-Vocabulary Instance Segmentation within 3D scenes. Objects within 3D environments exhibit diverse shapes, scales, and colors, making precise instance-level identification a challenging task. Recent advancements in Open-Vocabulary scene understanding have made significant strides in this area by employing class-agnostic 3D instance proposal networks for object localization and learning queryable features for each 3D mask. While these methods produce high-quality instance proposals, they struggle with identifying small-scale and geometrically ambiguous objects. The key idea of our method is a new module that aggregates 2D instance masks across frames and maps them to geometrically coherent point cloud regions as high-quality object proposals addressing the above limitations. These are then combined with 3D class-agnostic instance proposals to include a wide range of objects in the real world. To validate our approach, we conducted experiments on three prominent datasets, including ScanNet200, S3DIS, and Replica, demonstrating significant performance gains in segmenting objects with diverse categories over the state-of-the-art approaches.