MA Lee
yes, should work fine
It feels like there is a small mistake in your equation @kwea123: ``` P[0, 2] = 2 * (cx / W) - 1 P[1, 2] = 2 * (cy /...
Thank you for your inspiring question. Rendering using conical frustums as in MipNeRF and MipNeRF tries to solve the problem of multi-scale rendering (or anti-aliasing) instead of modeling blurriness. So...
We export the GT camera poses directly from the Blender file for the synthetic dataset (as is the cozyroom).
We extract from COLMAP (which is why it might be inaccurate and we don't use it for numerical evaluation)
It seems that the author tried to reproduce the modified canny detection but failed, so the original OpenCV's canny detection is used. So I find the bellow solution well approximated...
Paper title: 3D Video Loops from Asynchronous Input Paper link: https://arxiv.org/abs/2303.05312 Project page: https://limacv.github.io/VideoLoop3D_web/ Code link: https://github.com/limacv/VideoLoop3D This paper should be in a new category named ```新视点合成(Novel View Synthesis)```, which...
这个codebase没有考虑360输入的情况（比如far-field modeling，proposal sampline），所以如果需要跑在360的data中需要改NeRF的结构。一种（可能）更方便的解决办法是吧DSK model直接迁移到已有的360NeRF的codebase里。