# What You See is What You Get: Exploiting Visibility for 3D Object Detection

arXiv: https://arxiv.org/abs/1912.04986 · HTML: https://mlsanity.com/paper/1912.04986/

## Key facts

- Cited by 69 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 10 Dec 2019.

## Built on by (most cited first, 30 of 69)

- [Multimodal Virtual Point 3D Detection](https://mlsanity.com/paper/2111.06881/) (2021)
- [Offboard 3D Object Detection from Point Cloud Sequences](https://mlsanity.com/paper/2103.05073/) (2021)
- [RangeDet:In Defense of Range View for LiDAR-based 3D Object Detection](https://mlsanity.com/paper/2103.10039/) (2021)
- [3D Object Detection for Autonomous Driving: A Comprehensive Survey](https://mlsanity.com/paper/2206.09474/) (2022)
- [Panoptic-PolarNet: Proposal-free LiDAR Point Cloud Panoptic Segmentation](https://mlsanity.com/paper/2103.14962/) (2021)
- [3D Object Detection for Autonomous Driving: A Survey](https://mlsanity.com/paper/2106.10823/) (2021)
- [PillarNet: Real-Time and High-Performance Pillar-based 3D Object Detection](https://mlsanity.com/paper/2205.07403/) (2022)
- [FB-BEV: BEV Representation from Forward-Backward View Transformations](https://mlsanity.com/paper/2308.02236/) (2023)
- [FusionPainting: Multimodal Fusion with Adaptive Attention for 3D Object Detection](https://mlsanity.com/paper/2106.12449/) (2021)
- [AFDetV2: Rethinking the Necessity of the Second Stage for Object Detection from Point Clouds](https://mlsanity.com/paper/2112.09205/) (2021)
- [Differentiable Raycasting for Self-supervised Occupancy Forecasting](https://mlsanity.com/paper/2210.01917/) (2022)
- [M3DeTR: Multi-representation, Multi-scale, Mutual-relation 3D Object Detection with Transformers](https://mlsanity.com/paper/2104.11896/) (2021)
- [ALSO: Automotive Lidar Self-supervision by Occupancy estimation](https://mlsanity.com/paper/2212.05867/) (2022)
- [MSeg3D: Multi-modal 3D Semantic Segmentation for Autonomous Driving](https://mlsanity.com/paper/2303.08600/) (2023)
- [GrooMeD-NMS: Grouped Mathematically Differentiable NMS for Monocular 3D Object Detection](https://mlsanity.com/paper/2103.17202/) (2021)
- [Multi-Modal 3D Object Detection in Autonomous Driving: a Survey](https://mlsanity.com/paper/2106.12735/) (2021)
- [Panoptic-PHNet: Towards Real-Time and High-Precision LiDAR Panoptic Segmentation via Clustering Pseudo Heatmap](https://mlsanity.com/paper/2205.07002/) (2022)
- [Behind the Curtain: Learning Occluded Shapes for 3D Object Detection](https://mlsanity.com/paper/2112.02205/) (2021)
- [Beyond 3D Siamese Tracking: A Motion-Centric Paradigm for 3D Single Object Tracking in Point Clouds](https://mlsanity.com/paper/2203.01730/) (2022)
- [VISTA: Boosting 3D Object Detection via Dual Cross-VIew SpaTial Attention](https://mlsanity.com/paper/2203.09704/) (2022)
- [RangeLDM: Fast Realistic LiDAR Point Cloud Generation](https://mlsanity.com/paper/2403.10094/) (2024)
- [4D-Net for Learned Multi-Modal Alignment](https://mlsanity.com/paper/2109.01066/) (2021)
- [UnO: Unsupervised Occupancy Fields for Perception and Forecasting](https://mlsanity.com/paper/2406.08691/) (2024)
- [Towards Long-Tailed 3D Detection](https://mlsanity.com/paper/2211.08691/) (2022)
- [FastPillars: A Deployment-friendly Pillar-based 3D Detector](https://mlsanity.com/paper/2302.02367/) (2023)
- [Score refinement for confidence-based 3D multi-object tracking](https://mlsanity.com/paper/2107.04327/) (2021)
- [SA-BEV: Generating Semantic-Aware Bird's-Eye-View Feature for Multi-view 3D Object Detection](https://mlsanity.com/paper/2307.11477/) (2023)
- [Fooling LiDAR Perception via Adversarial Trajectory Perturbation](https://mlsanity.com/paper/2103.15326/) (2021)
- [Advancements in Point Cloud Data Augmentation for Deep Learning: A Survey](https://mlsanity.com/paper/2308.12113/) (2023)
- [Learning by Watching](https://mlsanity.com/paper/2106.05966/) (2021)

## Abstract (arXiv)

Recent advances in 3D sensing have created unique challenges for computer vision. One fundamental challenge is finding a good representation for 3D sensor data. Most popular representations (such as PointNet) are proposed in the context of processing truly 3D data (e.g. points sampled from mesh models), ignoring the fact that 3D sensored data such as a LiDAR sweep is in fact 2.5D. We argue that representing 2.5D data as collections of (x, y, z) points fundamentally destroys hidden information about freespace. In this paper, we demonstrate such knowledge can be efficiently recovered through 3D raycasting and readily incorporated into batch-based gradient learning. We describe a simple approach to augmenting voxel-based networks with visibility: we add a voxelized visibility map as an additional input stream. In addition, we show that visibility can be combined with two crucial modifications common to state-of-the-art 3D detectors: synthetic data augmentation of virtual objects and temporal aggregation of LiDAR sweeps over multiple time frames. On the NuScenes 3D detection benchmark, we show that, by adding an additional stream for visibility input, we can significantly improve the overall detection accuracy of a state-of-the-art 3D detector.