Height-Guided Projection Reparameterization for Camera-LiDAR Occupancy

TL;DR

This paper introduces HiPR, a novel camera-LiDAR occupancy framework that adaptively reparameterizes the projection space using height priors, improving scene understanding and outperforming existing methods.

Contribution

The paper proposes Height-Guided Projection Reparameterization (HiPR), which adaptively adjusts the projection space based on height priors to better capture scene geometry.

Findings

HiPR outperforms state-of-the-art methods in occupancy prediction.

The adaptive reparameterization improves feature aggregation accuracy.

The approach maintains real-time inference capabilities.

Abstract

3D occupancy prediction aims to infer dense, voxel-wise scene semantics from sensor observations, where the 2D-to-3D view transformation serves as a crucial step in bridging image features and volumetric representations. Most previous methods rely on a fixed projection space, where 3D reference points are uniformly sampled along pillars. However, such sampling struggles to capture the sparsity and height variations of real-world scenes, leading to ambiguous correspondences and unreliable feature aggregation. To address these challenges, we propose HiPR, a camera-LiDAR occupancy framework with Height-Guided Projection Reparameterization. HiPR first encodes LiDAR into a BEV height map to capture the maximum height of the point cloud. HiPR then adjusts the sampling range of each pillar using the height prior, enabling adaptive reparameterization of the projection space. As a result, the projected points are redistributed into geometrically meaningful regions rather than fixed ranges. Meanwhile, we mask out the invalid parts of the height map to avoid misleading the feature aggregation. In addition, to alleviate the training instability caused by noisy LiDAR-derived heights, we introduce a training-time Progressive Height Conditioning strategy, which gradually transitions the conditioning signal from ground-truth heights to LiDAR heights. Extensive experiments demonstrate that HiPR consistently outperforms existing state-of-the-art methods while maintaining real-time inference. The code and pretrained models can be found at https://github.com/yanzq95/HiPR.