jonbarron

Hello, this code assumes that you are providing a confidence image as input, and there is no code provided in this repo for computing a confidence image. The best choice...

The only constraint on the confidence is that it's non-negative. This code uses uint8 images but that's just a side effect of how image formats work. If you're getting all-black...

1) Yes, two resampling rounds works slightly better than one round. Sorry, we should have mentioned this in the paper as another difference. 2) This is discussed in the appendix...

Depends on what strategy you use for increasing the dimensionality of the IPE in mip-NeRF. If you use the same strategy, you should expect a similar improvement.

Yep, this code is designed to run on GPUs. If your GPU is unused when you run the code, you likely have installed JAX incorrectly. I'd take a look at...

It looks like you're running GHT on a list of points. Try running it on a histogram (or on a sorted list of points provided as input as "x", where...

Here you go: https://drive.google.com/file/d/1IevRDSI3bYGVhcZuDU9IZ81insF3w0XO/view?usp=sharing. I tuned the knobs a bit to give me 2.6. On Sun, Jul 26, 2020 at 2:42 PM Jon Barron wrote: > It looks like you're...

They're the hyperparameters of a scaled inverse chi-squared distribution, if you find that intuitive. The paper goes into much more detail on this but they kinda serve to let you...

Sorry, I have no idea how this would be done. Boosting algorithms are usually not agnostic to the choice of loss function.

Hi Federico, thanks for the kind words! I have no plans to write a matlab implementation of this code, but I'm happy to help support any effort you might have...