4 ms·
Pixel Objectness (2017)
- lainon 8y agoNote that there's a more recent paper than the one linked on the project page: https://arxiv.org/abs/1808.04702 https://arxiv.org/abs/1808.04702
- xfr 8y agoShould have (2017) in the title.
- dang 8y agoDone. Thanks!
- amelius 8y agoDid they cherry-pick the examples? It looks suspiciously good.
- fhood 8y agoThat's what I was wondering. This is way beyond the level of accuracy that I have seen from anything else.
- amelius 8y agoHonestly, I find this practice unacceptable, and unfortunately I see it a lot. The academic world should frown upon this kind of presentation.
- kevin_thibedeau 8y agoIt isn't clear from the paper but their appearance model shown in the video segmentation demo is likely the same thing. It shows poor performance at times (see the train at 0:32). http://vision.cs.utexas.edu/projects/fusionseg/ http://vision.cs.utexas.edu/projects/fusionseg/