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/

Global: Zendy Plus annual

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

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

3D instance segmentation methods often require fully-annotated dense labelsfor training, which are costly to obtain. In this paper, we present ClickSeg, anovel click-level weakly supervised 3D instance segmentation method thatrequires one point per instance annotation merely. Such a problem is verychallenging due to the extremely limited labels, which has rarely been solvedbefore. We first develop a baseline weakly-supervised training method, whichgenerates pseudo labels for unlabeled data by the model itself. To utilize theproperty of click-level annotation setting, we further propose a new trainingframework. Instead of directly using the model inference way, i.e., mean-shiftclustering, to generate the pseudo labels, we propose to use k-means with fixedinitial seeds: the annotated points. New similarity metrics are furtherdesigned for clustering. Experiments on ScanNetV2 and S3DIS datasets show thatthe proposed ClickSeg surpasses the previous best weakly supervised instancesegmentation result by a large margin (e.g., +9.4% mAP on ScanNetV2). Using0.02% supervision signals merely, ClickSeg achieves $\sim$90% of the accuracyof the fully-supervised counterpart. Meanwhile, it also achievesstate-of-the-art semantic segmentation results among weakly supervised methodsthat use the same annotation settings.

ClickSeg: 3D Instance Segmentation with Click-Level Weak Annotations