KITTI-360

Method

Simultaneous Diffusion (R2DM) [Simultaneous R2DM]
https://github.com/Ryan-Faulkner/Simultaneous-Diffusion-for-Pointclouds

Submitted on 4 Sep. 2025 13:30 by
Ryan Faulkner (University of Adelaide)

Running time:-
Environment:NVIDIA V100

Method Description:
Preprint is on Arxiv, paper has been accepted into the
26th International Conference on Digital Image
Computing: Techniques and Applications (DICTA 2025).

Multiple new LiDAR scans are synthesised using
diffusion, with the
original scan applied as a condition.

Furthermore, our novel applies each new scan as a
condition against the
others, enforcing consistency, and creating synthetic
points which are
more accurate (better geometric consistency), and also
less noisy.
Parameters:
parameters are as set in the code
Latex Bibtex:
@misc{faulkner2024simultaneousdiffusionsamplingcond
itional,
title={Simultaneous Diffusion Sampling for
Conditional LiDAR
Generation},
author={Ryan Faulkner and Luke Haub and Simon
Ratcliffe and Anh-
Dzung Doan and Ian Reid and Tat-Jun Chin},
year={2024},
eprint={2410.11628},
archivePrefix={arXiv},
primaryClass={cs.CV},
url={https://arxiv.org/abs/2410.11628},
}

Detailed Results

This page provides detailed results for the method(s) selected. For the first 5 test point clouds, we display the original image, the color-coded result and an error image. The error image contains 4 colors weighted by the confidence of the pseudo-ground truth:
red: the pixel has the wrong label and the wrong category
yellow: the pixel has the wrong label but the correct category
green: the pixel has the correct label
black: the groundtruth label is not used for evaluation

Test Set Average

Accuracy Completeness F1 mIoU class
80.37 29.49 43.15 3.88
This table as LaTeX

Test Image 0


Test Image 1


Test Image 2


Test Image 3


Test Image 4





eXTReMe Tracker