RefMask3D: Language-Guided Transformer for 3D Referring Segmentation

TL;DR

RefMask3D introduces a novel transformer-based model for 3D referring segmentation that effectively fuses vision and language features, achieving state-of-the-art results on multiple datasets.

Contribution

The paper proposes Geometry-Enhanced Group-Word Attention, Linguistic Primitives Construction, and an Object Cluster Module to improve multi-modal understanding in 3D segmentation.

Findings

Achieves new state-of-the-art performance on 3D referring segmentation datasets.

Outperforms previous methods by 3.16% mIoU on ScanRefer.

Effective multi-modal feature fusion for irregular point clouds.

Abstract

3D referring segmentation is an emerging and challenging vision-language task that aims to segment the object described by a natural language expression in a point cloud scene. The key challenge behind this task is vision-language feature fusion and alignment. In this work, we propose RefMask3D to explore the comprehensive multi-modal feature interaction and understanding. First, we propose a Geometry-Enhanced Group-Word Attention to integrate language with geometrically coherent sub-clouds through cross-modal group-word attention, which effectively addresses the challenges posed by the sparse and irregular nature of point clouds. Then, we introduce a Linguistic Primitives Construction to produce semantic primitives representing distinct semantic attributes, which greatly enhance the vision-language understanding at the decoding stage. Furthermore, we introduce an Object Cluster Module that analyzes the interrelationships among linguistic primitives to consolidate their insights and pinpoint common characteristics, helping to capture holistic information and enhance the precision of target identification. The proposed RefMask3D achieves new state-of-the-art performance on 3D referring segmentation, 3D visual grounding, and also 2D referring image segmentation. Especially, RefMask3D outperforms previous state-of-the-art method by a large margin of 3.16% mIoU} on the challenging ScanRefer dataset. Code is available at https://github.com/heshuting555/RefMask3D.