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
尊敬的作者,您好,感谢您的工作。当我复现您的工作在大场景数据集下,效果变的非常差,我发现渲染的深度图与法向图好像是错误的。没法在大场景下渲染正确的法向图与深度图,当我把法相优化关掉以后,效果就变得好起来了。
The text was updated successfully, but these errors were encountered:
我猜测可能是因为论文在公式(15)中用高斯中心点与相机主轴的夹角 去近似 射线高斯交点与相机主轴的夹角 去计算深度值,当高斯核比较小时这样近似是合理的,但当高斯核很大时,有可能会造成过大的误差,而大场景渲染应该包含很多大尺度的高斯核,故渲染的深度图可能效果不佳。可以试试TrimGS直接对大高斯分裂K个方法去缩小高斯核大小去渲染大场景。
Sorry, something went wrong.
您好,我更新了代码。目前代码不再使用中心去近似夹角。我在实验中还发现目前的分裂策略在大场景初始化噪声严重的时候表现较差,我之后也会将3DGS原始的分裂方法作为optional的选项加入代码。
请问这是改了哪部分的代码实现?
No branches or pull requests
尊敬的作者,您好,感谢您的工作。当我复现您的工作在大场景数据集下,效果变的非常差,我发现渲染的深度图与法向图好像是错误的。没法在大场景下渲染正确的法向图与深度图,当我把法相优化关掉以后,效果就变得好起来了。
The text was updated successfully, but these errors were encountered: