dvl-tum / defocus-net Goto Github PK

Hi, nice paper.
Thank you for releasing dataset link in readme.
However, there are only 500 scenes of training/testing set, which is a mismatch compared to what mentioned in the paper.
The all-in-focus ground truth and defocus maps are also missing.
Looking forward the complete dataset releasing.

hi, Do you have a testing script? If there is one I cannot find. Please help

Hello! You have a very interesting paper, and we would be happy to evaluate the results. When do you plan on publishing the evaluation code / training code / dataset?

due to blender version keep updating, when using the version 4.0 ,many api are changing ,some errors occured ,may I know what version is your blender ,and what platform ubuntu or windows,or do u have a new script fitting the newer blend version

Hi! Excuse me, this is a great paper! but I get an error when running run_training.py （[Errno 2] No such file or directory: '../models/a01_d02_t01/a01_d02_t01_ep1000.pth'）can you help me？

Thank you for sharing the code.

I am trying to run the training process. But when I try the command "python run_training.py", It report ERROR "FileNotFoundError: [Errno 2] No such file or directory: '../models/a01_d02_t01/a01_d02_t01_ep1000.pth'". So how can I get the pre-trained model?

Thank you for sharing the code.

I am trying to run the training process. But based on the default setting, it seems we are supposed to load a model from step1, and the training model is 2.

May I ask what they are? And where can I get the access to the MODEL1?

This is a very interesting work!
After running and modifying the code, I get the predicated depth map and focus map, but I cannot find a way to get the predicted all-in-focus image. Could you please give me a clue to help me get the all-in-focus output?
Thanks a lot!