Publication

Unsupervised Multi-Task Feature Learning on Point Clouds

AbstractWe introduce an unsupervised multi-task model to jointly learn point and shape features on point clouds. We define three unsupervised tasks including clustering, reconstruction, and self-supervised classification to train a multi-scale graph-based encoder. We evaluate our model on shape classification and segmentation benchmarks. The results suggest that it outperforms prior state-of-the-art unsupervised models: In the ModelNet40 classification task, it achieves an accuracy of 89.1% and in ShapeNet segmentation task, it achieves an mIoU of 68.2 and accuracy of 88.6%.

Download publication

Related Resources

See what’s new.

Publication

01/01/2011

Comet and Target Ghost: Techniques for Selecting Moving Targets

Numerous applications such as simulations, air traffic control…

Publication

01/01/1991

Issues in Combining Marking and Direct Manipulation Techniques

The direct manipulation paradigm has been effective in helping…

Publication

01/01/2021

Program Synthesis Guided Reinforcement Learning

A key challenge for reinforcement learning is solving long-horizon…

Article

12/06/2022

How the Construction Industry can Operate Sustainably: A New Way to Reduce Waste

Autodesk Research hosts a panel discussion: “Looking Towards a…

Get in touch

Something pique your interest? Get in touch if you’d like to learn more about Autodesk Research, our projects, people, and potential collaboration opportunities.

Contact us