gaussian-splatting

gaussian-splatting copied to clipboard

I encountered a problem when running "pip install submodules/diff-gaussian-rasterization". Help is urgently needed!

3

Processing d:\3dgs\gaussian-splatting\submodules\diff-gaussian-rasterization Preparing metadata (setup.py) ... done Building wheels for collected packages: diff_gaussian_rasterization Building wheel for diff_gaussian_rasterization (setup.py) ... error error: subprocess-exited-with-error × python setup.py bdist_wheel did not run successfully....

Rone4567

I've developed a custom rasterizer (replacing `submodules/diff-gaussian-rasterization`) and I would like to try viewing it using SIBR. How do I do this? Thank you!

kjy5

关于新代码为何效果比原论文好疑问？

2

在跑完新版代码的高斯代码后我发现指标会比原论文好很多，最后我发现是指令上的问题对需要下采样的数据集，通常指直接只用了-r=4指令，这会调用train 中的resized_image_rgb = PILtoTorch(image, resolution)进行下采样！而如果使用convert.py进行下采样的图片作为train.py的输入的话，其底层采用的是 ImageMagick 的 -resize 25%，两者似乎对最终渲染结果有一定的影响！在 resolution = round(orig_w/(resolution_scale * args.resolution)), round(orig_h/(resolution_scale * args.resolution))中也存在一个潜在的问题当使用bicycle数据集时候，使用-r=4后原先的宽度4146/4=1036.5,此时 round()函数会进行银行家舍入规则将0.5舍去，得到1036宽度的图像，这在后面的render.py,以及评估都是这个尺度。同样的如果在covert.py里面提前进行下采样得到的image_4里面是1037宽度的图像，这是因为ImageMagick会四舍五入！如果用image_4进行作为train的输入其后面的真实值和渲染都将是1037！最终问题的核心可能还是在mageMagick 的 -resize 25% PILtoTorch(image, resolution)的底层处理不同导致效果增加！

xiaohaiyo

The reproduced evaluated performance on tank dataset is much higher than the original paper reported.

2

Hi, I configured my environment according to the official repository and ran the “train” scene on the tank dataset. I ran it multiple times and the PSNR and SSIM values...

twowwj

background removal

13

I'm trying to separate the background from the foreground by making the background transparent. I tried making the background transparent on the input images before giving to gaussian-splatting but it...

mikebilly

hello，everyone！ I am wondering resize the tile size like 8x8,32x32 as many papers researched before. It seems that I only have to adjust the parameters BLOCK_X and BLOCK_Y, but soon...

layaout

Fail to instll SIBR_viewers in windows

2

Hi, I meet problem when I install the SIBR viewer: ` The downloadAndExtractZipFile didn't work...? -- Found a 3rdParty ASSIMP dir : E:/3DGS_Recurrence/gaussian-splatting/SIBR_viewers/extlibs/assimp. CMake Error at D:/cmake_3.30.5/share/cmake-3.30/Modules/FindPackageHandleStandardArgs.cmake:233 (message): ASSIMP wasn't...

curryxjh

Hello author, thank you very much for your excellent work. I would like to know your opinion on this issue.

weljam

RuntimeError: CUDA error: an illegal memory access was encountered

23

Hello, I was surprised by your work and tried to reproduce it with the code you've provided. However, every time I tried to run the code, it always failed to...

seohoiki3215

I have been little bit exploring the sources and noticed this redundant initialization :)

GalkinDenis