PointCloud Saliency Maps

Abstract

3D point-cloud recognition with PointNet and its variants has received remarkable progress. A missing ingredient, however, is the ability to automatically evaluate point-wise importance w.r.t.\! classification performance, which is usually reflected by a saliency map. A saliency map is an important tool as it allows one to perform further processes on point-cloud data. In this paper, we propose a novel way of characterizing critical points and segments to build point-cloud saliency maps. Our method assigns each point a score reflecting its contribution to the model-recognition loss. The saliency map explicitly explains which points are the key for model recognition. Furthermore, aggregations of highly-scored points indicate important segments/subsets in a point-cloud. Our motivation for constructing a saliency map is by point dropping, which is a non-differentiable operator. To overcome this issue, we approximate point-dropping with a differentiable procedure of shifting points towards the cloud centroid. Consequently, each saliency score can be efficiently measured by the corresponding gradient of the loss w.r.t the point under the spherical coordinates. Extensive evaluations on several state-of-the-art point-cloud recognition models, including PointNet, PointNet++ and DGCNN, demonstrate the veracity and generality of our proposed saliency map. Code for experiments is released on \url{https://github.com/tianzheng4/PointCloud-Saliency-Maps}.

arXiv commentAccepted to ICCV19 (oral)

References in corpus (10)

• Intriguing properties of neural networks
• Striving for Simplicity: The All Convolutional Net
• Obfuscated Gradients Give a False Sense of Security: Circumventing Defenses to Adversarial Examples
• Adversarial Machine Learning at Scale
• Dynamic Graph CNN for Learning on Point Clouds
• Provably Minimally-Distorted Adversarial Examples
• Boosting Adversarial Attacks with Momentum
• Curriculum Adversarial Training
• Generating 3D Adversarial Point Clouds
• Distributionally Adversarial Attack

Cited by in corpus (6)

• PointNetKL: Deep Inference for GICP Covariance Estimation in Bathymetric SLAM
• Extending Adversarial Attacks and Defenses to Deep 3D Point Cloud Classifiers
• DUP-Net: Denoiser and Upsampler Network for 3D Adversarial Point Clouds Defense
• On Isometry Robustness of Deep 3D Point Cloud Models under Adversarial Attacks
• Adversarial shape perturbations on 3D point clouds
• Verifiability and Predictability: Interpreting Utilities of Network Architectures for Point Cloud Processing