CPC G06V 20/64 (2022.01) [G06F 18/211 (2023.01); G06F 18/2163 (2023.01); G06F 18/22 (2023.01); G06F 18/23 (2023.01); G06F 18/251 (2023.01); G06T 9/001 (2013.01); G06V 10/26 (2022.01); G06V 10/806 (2022.01); G06V 10/82 (2022.01); G06V 20/70 (2022.01)] | 18 Claims |
1. A point cloud segmentation method, comprising:
encoding a to-be-processed point cloud to obtain a shared feature, the shared feature referring to a feature shared at a semantic level and at an instance level;
decoding the shared feature to obtain a semantic feature and an instance feature respectively;
adapting the semantic feature to an instance feature space and fusing the semantic feature with the instance feature, to obtain a semantic-fused instance feature of the point cloud, the semantic-fused instance feature representing an instance feature fused with the semantic feature;
dividing the semantic-fused instance feature of the point cloud by using an independent fully-connected layer of each point in the point cloud to process the semantic-fused instance feature of the point cloud, to obtain a semantic-fused instance feature of each point in the point cloud; and
determining an instance category to which each point belongs according to the semantic-fused instance feature of each point,
wherein the semantic feature is a semantic feature matrix, the instance feature is an instance feature matrix, and the adapting the semantic feature to an instance feature space and fusing the semantic feature with the instance feature, to obtain a semantic-fused instance feature of the point cloud comprises:
adapting the semantic feature matrix to the instance feature space by using an independent first fully connected layer of each point; and
performing element-wise addition on the semantic feature matrix adapted to the instance feature space and the instance feature matrix, to obtain a first matrix of the point cloud, the first matrix being a semantic-fused instance feature matrix.
|