[4eyes] FW: CVPR feedback

[Matthew Turk]

As an addition to last week’s CVPR discussion at the lab meeting, here are some additional comments on interesting papers from Jedrzej (some of which were mentioned last week):
 
 
There was no huge breakthrough paper at this CVPR, but here are just couple papers I found interesting and worth recommending:
*	http://openaccess.thecvf.com/content_cvpr_2018/papers/Zamir_Taskonomy_Disentangling_Task_CVPR_2018_paper.pdf   (Best Paper Award) - on transfer learning between different visual tasks
*	http://openaccess.thecvf.com/content_cvpr_2018/papers/Kundu_3D-RCNN_Instance-Level_3D_CVPR_2018_paper.pdf    use of rendering for better 3D instance-segmentation
*	http://openaccess.thecvf.com/content_cvpr_2018/papers/Qi_Low-Shot_Learning_With_CVPR_2018_paper.pdf    transfer learning for low-shot image classification
*	http://openaccess.thecvf.com/content_cvpr_2018/papers/Wang_Low-Shot_Learning_From_CVPR_2018_paper.pdf    low-shot learning by 'hallucinating' additional samples
I would say first two papers were big in general and the latter two I found important in terms of closeness to our approaches and ideas.
There are many more papers of course, but I would say those peaked my curiosity the most.
 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cs.ucsb.edu/pipermail/ilab-users/attachments/20180802/86e405cc/attachment-0001.html>