ingra14m / floater-free-gaussian-splatting Goto Github PK

Can you briefly tell me what code you changed?

absgs seems pretty straightforward

Thanks

The metric numbers reported in this README is really impressive. I am curious how you prepare the input dataset, are you using the convert.py script to create downscale 4 images?

I found most papers reported their numbers based on the prepared version in mipnerf360 dataset.

Hello, thanks for sharing your excellent work. I'm quite new to the CUDA code and I'm wondering if you could point out to me which part of the CUDA code in submodules/diff-depth-diff-gaussian-rasterization you modified to calculate the homodirectional view-space positional gradient in the absGaussian paper? Thanks in advance.

When ran the render.py to get a video, I have got this error.

Traceback (most recent call last):
File "render.py", line 264, in
render_sets(model.extract(args), args.iteration, pipeline.extract(args), args.skip_train, args.skip_test, args.render_video)
File "render.py", line 245, in render_sets
render_video(dataset.model_path, scene.loaded_iter, scene.getTrainCameras(), gaussians, pipeline, background)
TypeError: 'bool' object is not callable