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

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.

PDF Abstract
Task Dataset Model Metric Name Metric Value Global Rank Benchmark
3D Open-Vocabulary Instance Segmentation Replica Open3DIS mAP 18.1 # 1
3D Open-Vocabulary Instance Segmentation S3DIS Open3DIS AP50 Base B8/N4 60.8 # 1
AP50 Novel B8/N4 26.3 # 2
AP50 Base B6/N6 50.0 # 2
AP50 Novel B6/N6 29.0 # 2
3D Instance Segmentation ScanNet200 Open3DIS (Open-Vocabulary) mAP 23.7 # 4
3D Open-Vocabulary Instance Segmentation ScanNet200 Open3DIS mAP 23.7 # 1
AP50 29.4 # 1
AP25 32.8 # 1
AP Head 27.8 # 1
AP Common 21.2 # 1
AP Tail 21.8 # 1

Methods


No methods listed for this paper. Add relevant methods here