We read every piece of feedback, and take your input very seriously.
To see all available qualifiers, see our documentation.
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
我最近也在复现这篇。 如论文所说只有第一帧是用noise_frame的LL的sigma作初始化 所以每次循环帧训练的时候,
EMVD/structure.py
Line 220 in 975a2f4
The text was updated successfully, but these errors were encountered:
如EMVD/structure.py#L221和L222所示,DCNN模块的sigma由当前帧和融合帧计算所得,即 sigma = (1 - gamma) * (1 - gamma) * sigma_ll0 + gamma * gamma * sigma_ll1 denoise_in = torch.cat([fusion_out, ll1, sigma], dim=1), 符合论文中公式(9)所述。
sigma = (1 - gamma) * (1 - gamma) * sigma_ll0 + gamma * gamma * sigma_ll1
denoise_in = torch.cat([fusion_out, ll1, sigma], dim=1)
Sorry, something went wrong.
我知道,公式(9)是符合的,我是说sigma_ll0只有第一帧需要通过a*z_LL+b算,后面帧的sigma_ll0是前面帧传入的
我觉得作者实现的没有问题啊,sigma_ll0 是基于ll0计算,而ll0 来自前一帧的。求大佬赐教。
No branches or pull requests
我最近也在复现这篇。
![image](https://user-images.githubusercontent.com/42491565/127636049-2294d312-733e-45ff-8942-1ab51ffa6f4d.png)
如论文所说只有第一帧是用noise_frame的LL的sigma作初始化
所以每次循环帧训练的时候,
EMVD/structure.py
Line 220 in 975a2f4
这里只有“第一帧”是这样算的,后面的帧应该是由前面帧算的结果传入的
The text was updated successfully, but these errors were encountered: