ICM-3D: Instantiated Category Modeling for 3D Instance Segmentation

Separating 3D point clouds into individual instances is an important task for vision. It challenging due to the unknown and varying number of in a scene. Existing deep learning based works focus on two-step pipeline: first learn feature embedding then cluster points. Such pipeline leads disconnected intermediate objectives. In this paper, we propose integrated reformulation instance segmentation as per-point classification problem. We ICM-3D, single-step method segment via instantiated categorization. The augmented category information automatically constructed from spatial positions. conduct extensive experiments verify effectiveness ICM-3D show that it obtains inspiring performance across multiple frameworks, backbones benchmarks.