您好，作者，有疑问想请教一下

[thissbb]

请问为什么要与gt做这个，有什么好处吗？这样做有理论依据吗？
mean_restored =cv2.cvtColor(restored.astype(np.float32),cV2.COLOR BGR2GRAY).mean()
mean target = cv2.cvtColor(target.astype(np.float32)，cV2.COLOR BGR2GRAY).mean()
restored =np.clip(restored*(mean target/meanrestored)，0，1)
期待您的回复！感谢！

[movingforward100]

We are just following Kind and LLFlow, please refer to https://github.com/wyf0912/LLFlow/blob/main/code/test.py . If you find this repo useful, please give it a star ⭐ and consider citing our paper in your research. Thank you.