search

princeton-vl / RAFT

BSD 3-Clause "New" or "Revised" License
3.25k stars 632 forks source link

Question about the flow estimation. #4

Closed MarioDuran closed 4 years ago

MarioDuran commented 4 years ago

Hi! first at all, thank you so much for sharing this work, is really amazing!

I´m using this repo to generate an optic flow of a generical video, we are working on face anti spoofing techniques and we want to use an approach like this, but Im getting weird results which I do not know if they are normal, you can see the output here. https://drive.google.com/file/d/1-ALLjgPthC4P52F-V35dVDkTZq2EbQyN/view?usp=sharing

is this the normal output of this framework? it seems a little noisy, but I do not know if I messed up somewhere.

this is the arguments I´m using, thanks in advance for any advice or suggestion!

parser.add_argument('--model', default = 'models/chairs+things.pth', help="restore checkpoint") parser.add_argument('--small', action='store_true', help='use small model') parser.add_argument('--iters', type=int, default=50)

zachteed commented 4 years ago

Hi Mario,

I'm not able to open the video you attached. Could you possibly provide a link in different format or include a link to the original video?

MarioDuran commented 4 years ago

Hi! thanks again for your help! I uploaded the video to youtube, hope it helps to visualize it better. https://youtu.be/qoSbfaX1PMA

thank you so much in advance for your time!

kwea123 commented 4 years ago

Some frames are good, some are noisy... If it's not due to bugs, I suspect that it's in some sort of adversarial attack. Can you also share the original color video?