fgabel / deep-reinforcement-learning-for-visual-object-tracking-in-videos

paper implementation for the Machine Learning for Computer Vision lecture

Deep Reinforcement Learning for Visual Object Tracking in Videos

deep-reinforcement-learning-for-visual-object-tracking-in-videos's Issues

Have a look at the dataset

Check the videos for squared boxes

Is this code complete? How to train and test it?

Check how to/Convert Yolo to Pytorch

See title - possibilites
https://github.com/explosion/lightnet

More probable I guess; Haver closer look at "partial.py"
https://github.com/marvis
https://github.com/marvis/pytorch-yolo2

or this
https://github.com/makora9143/yolo-pytorch

Rewrite everything to 4 coordinates

drawing, loss function and more probably

How to do BackPropogation ?

Can you please explain how have you implemented the weight updation part?. I am confused on how to calculate the Ut. If you can't share the code can you please give me some pointers to implement it?.

Further, I'll be happy to collaborate in implementing this paper.

Stuff to be done on the server

fix errors on the file forward_test on the server:
"image dimensions must agree"

can be reproduced by running "python forward_test.py" in bundgaard_gabel/drlt/

Dataloader - T sequential bilder

Dataloader - T sequential bilder aus einem Video werden geladen

Rescale Pictures to [0,1]

E-Mail Sven for server access

See title

test

blabla

Implement Loss functions

Implement the loss functions mentioned in the paper in utils.py