search

Neooolee / WHUS2-CD

This is a cloud detection validation dataset for Sentinel-2A images
13 stars 5 forks source link

Steps to implement code #2

Open NakitaOza opened 2 years ago

NakitaOza commented 2 years ago

Hi! I read your paper and the results are very promising! I am currently pursuing my master thesis in the same field of cloud detection. I see that you have uploaded some training code in this repository. I would like to reproduce these results. I did not find any evaluation/testing function. Is there some where wherein I can input images to the 'trained' algorithm and see the results of masking on the given input files?

Neooolee commented 2 years ago

The evaluation/testing process is the same as other semantic segmentation algorithms such as U-Net. Just need to replace the algorithm code (one line) with our code and loading the pre-trained model. Then run the evaluation/testing code as you do with U-Net.

ludongxing commented 1 year ago

hello! Is there any detailed description of the procedure execution order？?please

Neooolee commented 1 year ago

Just replace the file path in Autoprocess.py, then run Autoprocess.py.