Continual Learning for Pose-Agnostic Object Recognition in 3D Point Clouds

TL;DR

This paper introduces a novel continual learning model for 3D point cloud object recognition that is pose-agnostic, leveraging geometric equivariance to handle dynamic and unpredictable object poses effectively.

Contribution

The work proposes a new continual learning approach that incorporates geometric equivariance as prior knowledge, addressing pose variability without increasing data augmentation complexity.

Findings

Outperforms existing methods on mainstream point cloud datasets.

Effectively handles pose variability in continual learning scenarios.

Ablation studies validate each component's contribution.

Abstract

Continual Learning aims to learn multiple incoming new tasks continually, and to keep the performance of learned tasks at a consistent level. However, existing research on continual learning assumes the pose of the object is pre-defined and well-aligned. For practical application, this work focuses on pose-agnostic continual learning tasks, where the object's pose changes dynamically and unpredictably. The point cloud augmentation adopted from past approaches would sharply rise with the task increment in the continual learning process. To address this problem, we inject the equivariance as the additional prior knowledge into the networks. We proposed a novel continual learning model that effectively distillates previous tasks' geometric equivariance information. The experiments show that our method overcomes the challenge of pose-agnostic scenarios in several mainstream point cloud datasets. We further conduct ablation studies to evaluate the validation of each component of our approach.