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

We aim to improve the performance of regressing hand keypoints and segmentingpixel-level hand masks under new imaging conditions (e.g., outdoors) when weonly have labeled images taken under very different conditions (e.g., indoors).In the real world, it is important that the model trained for both tasks worksunder various imaging conditions. However, their variation covered by existinglabeled hand datasets is limited. Thus, it is necessary to adapt the modeltrained on the labeled images (source) to unlabeled images (target) with unseenimaging conditions. While self-training domain adaptation methods (i.e.,learning from the unlabeled target images in a self-supervised manner) havebeen developed for both tasks, their training may degrade performance when thepredictions on the target images are noisy. To avoid this, it is crucial toassign a low importance (confidence) weight to the noisy predictions duringself-training. In this paper, we propose to utilize the divergence of twopredictions to estimate the confidence of the target image for both tasks.These predictions are given from two separate networks, and their divergencehelps identify the noisy predictions. To integrate our proposed confidenceestimation into self-training, we propose a teacher-student framework where thetwo networks (teachers) provide supervision to a network (student) forself-training, and the teachers are learned from the student by knowledgedistillation. Our experiments show its superiority over state-of-the-artmethods in adaptation settings with different lighting, grasping objects,backgrounds, and camera viewpoints. Our method improves by 4% the multi-taskscore on HO3D compared to the latest adversarial adaptation method. We alsovalidate our method on Ego4D, egocentric videos with rapid changes in imagingconditions outdoors.

Domain Adaptive Hand Keypoint and Pixel Localization in the Wild