We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:


Current browse context:


Change to browse by:

References & Citations

DBLP - CS Bibliography


(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Computer Vision and Pattern Recognition

Title: Investigate Indistinguishable Points in Semantic Segmentation of 3D Point Cloud

Abstract: This paper investigates the indistinguishable points (difficult to predict label) in semantic segmentation for large-scale 3D point clouds. The indistinguishable points consist of those located in complex boundary, points with similar local textures but different categories, and points in isolate small hard areas, which largely harm the performance of 3D semantic segmentation. To address this challenge, we propose a novel Indistinguishable Area Focalization Network (IAF-Net), which selects indistinguishable points adaptively by utilizing the hierarchical semantic features and enhances fine-grained features for points especially those indistinguishable points. We also introduce multi-stage loss to improve the feature representation in a progressive way. Moreover, in order to analyze the segmentation performances of indistinguishable areas, we propose a new evaluation metric called Indistinguishable Points Based Metric (IPBM). Our IAF-Net achieves the comparable results with state-of-the-art performance on several popular 3D point cloud datasets e.g. S3DIS and ScanNet, and clearly outperforms other methods on IPBM.
Comments: AAAI2021
Subjects: Computer Vision and Pattern Recognition (cs.CV); Image and Video Processing (eess.IV)
Cite as: arXiv:2103.10339 [cs.CV]
  (or arXiv:2103.10339v4 [cs.CV] for this version)

Submission history

From: Junhao Zhang [view email]
[v1] Thu, 18 Mar 2021 15:54:59 GMT (17147kb,D)
[v2] Tue, 24 Aug 2021 11:54:40 GMT (44844kb,D)
[v3] Wed, 25 Aug 2021 11:09:08 GMT (44844kb,D)
[v4] Thu, 26 Aug 2021 03:11:01 GMT (41631kb,D)

Link back to: arXiv, form interface, contact.