# Explaining Deep Neural Networks for Point Clouds using Gradient-based Visualisations

arXiv: https://arxiv.org/abs/2207.12984 · HTML: https://mlsanity.com/paper/2207.12984/

## Key facts

- Cited by 3 papers in the map (data through 14 Oct 2026).
- Builds on 20 of its 33 references that are in the map.
- First posted on arXiv 26 Jul 2022.

## Builds on (most cited first, 20 of 20)

- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [Microsoft COCO: Common Objects in Context](https://mlsanity.com/paper/1405.0312/) (2014)
- [PointNet: Deep Learning on Point Sets for 3D Classification and Segmentation](https://mlsanity.com/paper/1612.00593/) (2016)
- [PointNet++: Deep Hierarchical Feature Learning on Point Sets in a Metric Space](https://mlsanity.com/paper/1706.02413/) (2017)
- [Learning Deep Features for Discriminative Localization](https://mlsanity.com/paper/1512.04150/) (2015)
- [Deep Inside Convolutional Networks: Visualising Image Classification Models and Saliency Maps](https://mlsanity.com/paper/1312.6034/) (2013)
- [Visualizing and Understanding Convolutional Networks](https://mlsanity.com/paper/1311.2901/) (2013)
- [Dynamic Graph CNN for Learning on Point Clouds](https://mlsanity.com/paper/1801.07829/) (2018)
- [Sanity Checks for Saliency Maps](https://mlsanity.com/paper/1810.03292/) (2018)
- [RISE: Randomized Input Sampling for Explanation of Black-box Models](https://mlsanity.com/paper/1806.07421/) (2018)
- [Deep Hough Voting for 3D Object Detection in Point Clouds](https://mlsanity.com/paper/1904.09664/) (2019)
- [Layer-wise Relevance Propagation for Neural Networks with Local Renormalization Layers](https://mlsanity.com/paper/1604.00825/) (2016)
- [SPLATNet: Sparse Lattice Networks for Point Cloud Processing](https://mlsanity.com/paper/1802.08275/) (2018)
- [PointCloud Saliency Maps](https://mlsanity.com/paper/1812.01687/) (2018)
- [RL-GAN-Net: A Reinforcement Learning Agent Controlled GAN Network for Real-Time Point Cloud Shape Completion](https://mlsanity.com/paper/1904.12304/) (2019)
- [3D ShapeNets: A Deep Representation for Volumetric Shapes](https://mlsanity.com/paper/1406.5670/) (2014)
- [Learning Representations and Generative Models for 3D Point Clouds](https://mlsanity.com/paper/1707.02392/) (2017)
- [From Image-level to Pixel-level Labeling with Convolutional Networks](https://mlsanity.com/paper/1411.6228/) (2014)
- [Visualizing Point Cloud Classifiers by Curvature Smoothing](https://mlsanity.com/paper/1911.10415/) (2019)
- [Geometric Back-projection Network for Point Cloud Classification](https://mlsanity.com/paper/1911.12885/) (2019)

## Built on by (most cited first, 3 of 3)

- [ExpPoint-MAE: Better interpretability and performance for self-supervised point cloud transformers](https://mlsanity.com/paper/2306.10798/) (2023)
- [Interpretable Affordance Detection on 3D Point Clouds with Probabilistic Prototypes](https://mlsanity.com/paper/2504.18355/) (2025)
- [Explainable LiDAR 3D Point Cloud Segmentation and Clustering for Detecting Airplane-Generated Wind Turbulence](https://mlsanity.com/paper/2503.00518/) (2025)

## Abstract (arXiv)

Explaining decisions made by deep neural networks is a rapidly advancing research topic. In recent years, several approaches have attempted to provide visual explanations of decisions made by neural networks designed for structured 2D image input data. In this paper, we propose a novel approach to generate coarse visual explanations of networks designed to classify unstructured 3D data, namely point clouds. Our method uses gradients flowing back to the final feature map layers and maps these values as contributions of the corresponding points in the input point cloud. Due to dimensionality disagreement and lack of spatial consistency between input points and final feature maps, our approach combines gradients with points dropping to compute explanations of different parts of the point cloud iteratively. The generality of our approach is tested on various point cloud classification networks, including 'single object' networks PointNet, PointNet++, DGCNN, and a 'scene' network VoteNet. Our method generates symmetric explanation maps that highlight important regions and provide insight into the decision-making process of network architectures. We perform an exhaustive evaluation of trust and interpretability of our explanation method against comparative approaches using quantitative, quantitative and human studies. All our code is implemented in PyTorch and will be made publicly available.