2021 IEEE/CVF International Conference on Computer Vision (ICCV)· 2021
Point Transformer
- 2,219citations
- 2021year
Short summary
A novel 'Point Transformer' architecture leverages self-attention mechanisms for 3D point cloud processing, achieving state-of-the-art results in tasks like semantic scene segmentation.
AI-generated from the title and abstract; the full text is not read.
Key points
- Introduces the Point Transformer, a self-attention-based network for 3D point cloud processing.
- Achieves 70.4% mIoU on the S3DIS dataset for semantic scene segmentation.
- Outperforms the strongest prior model by 3.3 absolute percentage points.
- First model to surpass the 70% mIoU threshold on this task.
AI-generated from the title and abstract; the full text is not read.
Abstract
Self-attention networks have revolutionized natural language processing and are making impressive strides in image analysis tasks such as image classification and object detection. Inspired by this success, we investigate the application of self-attention networks to 3D point cloud processing. We design self-attention layers for point clouds and use these to construct self-attention networks for tasks such as semantic scene segmentation, object part segmentation, and object classification. Our Point Transformer design improves upon prior work across domains and tasks. For example, on the challenging S3DIS dataset for large-scale semantic scene segmentation, the Point Transformer attains an mIoU of 70.4% on Area 5, outperforming the strongest prior model by 3.3 absolute percentage points and crossing the 70% mIoU threshold for the first time.
The authors' abstract, as published at the source. 2021 IEEE/CVF International Conference on Computer Vision (ICCV), 2021 · DOI ↗
Continue with a free account
Ask the paper: 3 free questions a day about this paper; save it, get its citation, new summaries every day for your field. Takeaways are Premium.
Continue free on the webSign in with Google or Apple; no card needed. You come back to this paper.
On your phone:
Field: Computational Mechanics
Computational MechanicsEngineering