English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Zendy Plus

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Zendy Tools

Zendy Open

Collecting and labeling the registered 3D point cloud is costly. As a result,3D resources for training are typically limited in quantity compared to the 2Dimages counterpart. In this work, we deal with the data scarcity challenge of3D tasks by transferring knowledge from strong 2D models via RGB-D images.Specifically, we utilize a strong and well-trained semantic segmentation modelfor 2D images to augment RGB-D images with pseudo-label. The augmented datasetcan then be used to pre-train 3D models. Finally, by simply fine-tuning on afew labeled 3D instances, our method already outperforms existingstate-of-the-art that is tailored for 3D label efficiency. We also show thatthe results of mean-teacher and entropy minimization can be improved by ourpre-training, suggesting that the transferred knowledge is helpful insemi-supervised setting. We verify the effectiveness of our approach on twopopular 3D models and three different tasks. On ScanNet official evaluation, weestablish new state-of-the-art semantic segmentation results on thedata-efficient track.

Data Efficient 3D Learner via Knowledge Transferred from 2D Model