search

YanchaoYang / FDA

Fourier Domain Adaptation for Semantic Segmentation
491 stars 79 forks source link

looks promising, would love to try out your code! #1

Closed zoharbarzelay closed 4 years ago

zoharbarzelay commented 4 years ago

Hi,

Pretty cool work! Any updates on when the code would be released?

Thanks! Z.

gsygsy96 commented 4 years ago

Me too

kjjeong78 commented 4 years ago

I have done the FDA with python. However, there are some artifact because we transform the each channel r, g, b. How do you remove that?

image

dongyangcai commented 4 years ago

looking forward to your code!

YanchaoYang commented 4 years ago

Guys, thanks for your interest. I have just uploaded the source code and the trained models. Please have fun!

YanchaoYang commented 4 years ago

I have done the FDA with python. However, there are some artifact because we transform the each channel r, g, b. How do you remove that?

image

Did you do any normalization or mean subtraction before FDA? If so, you may want to remove these preprocessings.

Also, I have released the python version of FDA, and its usage in FDA_demo.py. Could you take a look and see if this will be useful?

kjjeong78 commented 4 years ago

I have done the FDA with python. However, there are some artifact because we transform the each channel r, g, b. How do you remove that? image

Did you do any normalization or mean subtraction before FDA? If so, you may want to remove these preprocessings.

Also, I have released the python version of FDA, and its usage in FDA_demo.py. Could you take a look and see if this will be useful?

Thank you. There are no preprocessings. I will try to test with your code and check my code what is a cause.

zoharbarzelay commented 4 years ago

thanks!!