CaRe-Ego: Contact-aware Relationship Modeling for Egocentric Interactive Hand-object Segmentation
Yuejiao Su, Yi Wang, and Lap-Pui Chau
Comparison results on the EgoHOS in-domain test set.
Comparison results on the EgoHOS out-of-domain test set.
Comparison results on the out-of-distribution mini-HOI4D dataset. The dataset of mini-HOI4D will be released soon.
Comparison results on the EgoHOS in-domain test set measured by IoU/Acc and mIoU/mAcc.
Comparison results on the EgoHOS out-of-domain test set measured by IoU/Acc and mIoU/mAcc.
Comparison results on the mini-HOI4D test set measured by IoU/Acc and mIoU/mAcc.
Although the CaRe-Ego is performed on Egocentric images, we can validate it on out-of-distribution videos frame-by-frame.
The research work was conducted in the JC STEM Lab of Machine Learning and Computer Vision funded by The Hong Kong Jockey Club Charities Trust.
The code of the CaRe-Ego is built upon the MMsegmentation codebase, thanks for their work.