LiNeXt: Revisiting LiDAR Completion with Efficient Non-Diffusion Architectures

2025-11-13

AI Summary

To address the slow, non-real-time inference of diffusion-based 3D LiDAR scene completion in autonomous driving—caused by iterative sampling—this paper proposes LiNeXt, an efficient non-diffusion architecture. Methodologically, LiNeXt introduces three key innovations: (1) a single-step Noise-to-Coarse (N2C) denoising module that replaces iterative diffusion; (2) a distance-aware Selected Repeat strategy to accommodate the near-dense/far-sparse point cloud distribution, enabling uniform structural completion; and (3) a lightweight Refine module for enhanced geometric detail fidelity. Evaluated on SemanticKITTI, LiNeXt achieves a 199.8× inference speedup over LiDiff, reduces Chamfer Distance by 50.7%, and uses only 6.1% of LiDiff’s parameters—demonstrating a substantial improvement in the efficiency–accuracy trade-off for real-time 3D scene understanding.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Search and Optimization: Learning to Search
• Natural Language Processing: Sentence-level Semantics, Textual Inference, etc.

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Economics, Online Markets and Human Computation: Architectures and workflows that use LLMs for crowd work
• Search and Retrieval-Augmented AI: Large language models for search

Abstract

3D LiDAR scene completion from point clouds is a fundamental component of perception systems in autonomous vehicles. Previous methods have predominantly employed diffusion models for high-fidelity reconstruction. However, their multi-step iterative sampling incurs significant computational overhead, limiting its real-time applicability. To address this, we propose LiNeXt-a lightweight, non-diffusion network optimized for rapid and accurate point cloud completion. Specifically, LiNeXt first applies the Noise-to-Coarse (N2C) Module to denoise the input noisy point cloud in a single pass, thereby obviating the multi-step iterative sampling of diffusion-based methods. The Refine Module then takes the coarse point cloud and its intermediate features from the N2C Module to perform more precise refinement, further enhancing structural completeness. Furthermore, we observe that LiDAR point clouds exhibit a distance-dependent spatial distribution, being densely sampled at proximal ranges and sparsely sampled at distal ranges. Accordingly, we propose the Distance-aware Selected Repeat strategy to generate a more uniformly distributed noisy point cloud. On the SemanticKITTI dataset, LiNeXt achieves a 199.8x speedup in inference, reduces Chamfer Distance by 50.7%, and uses only 6.1% of the parameters compared with LiDiff. These results demonstrate the superior efficiency and effectiveness of LiNeXt for real-time scene completion.

Problem

• Efficient LiDAR scene completion without diffusion models' computational overhead
• Single-pass denoising and refinement for real-time point cloud completion
• Addressing distance-dependent sparsity in LiDAR data distribution

Innovation

• Lightweight non-diffusion network for fast point cloud completion
• Single-pass denoising module eliminating iterative sampling steps
• Distance-aware strategy generating uniformly distributed noisy inputs