search

GuoLanqing / ShadowFormer

ShadowFormer (AAAI2023), Pytorch implementation
MIT License
141 stars 16 forks source link

RMSE or MAE #33

Open Amy-214 opened 8 months ago

Amy-214 commented 8 months ago

May I ask the author, why is the RMSE written in the evaluation index in the paper, but the MAE is calculated on the evaluation code?

KasuganoLove commented 7 months ago

May I ask the author, why is the RMSE written in the evaluation index in the paper, but the MAE is calculated on the evaluation code?

I have the same question. Note that in SID the paper mentioned this in README.md. (https://github.com/cvlab-stonybrook/SID)

Note on the shadow removal evaluation code: We recently figured that the RMSE evaluation code that many papers have been using (including ours) is actually calculating Mean Absolute Error. We will retrospectively fix this in all our papers and suggest everyone to do the same.

Is the RMSE in this paper the same?