Transductive Zero-Shot Learning for 3D Point Cloud Classification

TL;DR

This paper introduces a novel transductive zero-shot learning method for 3D point cloud classification, utilizing a new triplet loss and demonstrating state-of-the-art results and cross-domain applicability.

Contribution

It extends transductive ZSL and GZSL to 3D point clouds with a novel triplet loss, also applicable to 2D images, and establishes new benchmarks.

Findings

Achieves state-of-the-art ZSL and GZSL performance in 3D point cloud classification.

Demonstrates the method's effectiveness on 2D image classification.

Introduces a novel triplet loss leveraging unlabeled test data.

Abstract

Zero-shot learning, the task of learning to recognize new classes not seen during training, has received considerable attention in the case of 2D image classification. However despite the increasing ubiquity of 3D sensors, the corresponding 3D point cloud classification problem has not been meaningfully explored and introduces new challenges. This paper extends, for the first time, transductive Zero-Shot Learning (ZSL) and Generalized Zero-Shot Learning (GZSL) approaches to the domain of 3D point cloud classification. To this end, a novel triplet loss is developed that takes advantage of unlabeled test data. While designed for the task of 3D point cloud classification, the method is also shown to be applicable to the more common use-case of 2D image classification. An extensive set of experiments is carried out, establishing state-of-the-art for ZSL and GZSL in the 3D point cloud domain, as well as demonstrating the applicability of the approach to the image domain.