LittleMount / DescatterNet-for-unseen-real-world-objects

In this study, we propose a deep-learning-based method to image through dynamic scattering media in a non-invasive manner under incoherence illumination and obtain superior restoration result.
4 stars 5 forks source link

'unet_use_in_hardware.py' file not provided #3

Open momo773510 opened 1 month ago

momo773510 commented 1 month ago

In file main_realshow the unet_use_in_hardware and some functions cannot be resolved. According to the structure of code the CLANE algo missing, and I wonder if the file where the funcs defined are not provided in github, or perhaps I use the mismatching edition of opencv or other packages? Thanks. 屏幕截图 2024-09-18 111753

Python version: 3.11.9 Tensorflow & keras: 2.15 opencv-python: 4.10.0.84 scikit-image: 0.24.0 scipy: 1.14.1 torch: 2.4.1

momo773510 commented 1 month ago

And it seems in main_realshow.py the UNet is implemented with torch, while in main_unet.py there is a tensorflow model. The code annotation shows the transfer from tf to torch. Hope a updated edition.