PCAN: 3D Attention Map Learning Using Contextual Information for Point Cloud Based Retrieval
This addresses the challenge of efficient feature encoding for place recognition in vision, but it appears incremental as it builds on existing attention mechanisms for point clouds.
The paper tackles the problem of encoding local features into a discriminative global descriptor for point cloud based retrieval in place recognition, and the result is that the proposed PCAN network outperforms current state-of-the-art approaches on various benchmark datasets.
Point cloud based retrieval for place recognition is an emerging problem in vision field. The main challenge is how to find an efficient way to encode the local features into a discriminative global descriptor. In this paper, we propose a Point Contextual Attention Network (PCAN), which can predict the significance of each local point feature based on point context. Our network makes it possible to pay more attention to the task-relevent features when aggregating local features. Experiments on various benchmark datasets show that the proposed network can provide outperformance than current state-of-the-art approaches.