Exploring Hierarchical Graph Representation for Large-Scale Zero-Shot Image Classification

TL;DR

This paper introduces HGR-Net, a hierarchical graph-based framework that significantly improves zero-shot image classification on large-scale datasets like ImageNet-21K by leveraging class inheritance relations.

Contribution

The paper presents a novel hierarchical graph knowledge representation framework for zero-shot learning, outperforming existing methods on large-scale benchmarks.

Findings

HGR-Net outperforms all existing techniques on ImageNet-21K by 7%.

HGR-Net is effective in few-shot learning scenarios.

The method generalizes well to smaller datasets like ImageNet-21K-P.

Abstract

The main question we address in this paper is how to scale up visual recognition of unseen classes, also known as zero-shot learning, to tens of thousands of categories as in the ImageNet-21K benchmark. At this scale, especially with many fine-grained categories included in ImageNet-21K, it is critical to learn quality visual semantic representations that are discriminative enough to recognize unseen classes and distinguish them from seen ones. We propose a \emph{H}ierarchical \emph{G}raphical knowledge \emph{R}epresentation framework for the confidence-based classification method, dubbed as HGR-Net. Our experimental results demonstrate that HGR-Net can grasp class inheritance relations by utilizing hierarchical conceptual knowledge. Our method significantly outperformed all existing techniques, boosting the performance by 7\% compared to the runner-up approach on the ImageNet-21K benchmark. We show that HGR-Net is learning-efficient in few-shot scenarios. We also analyzed our method on smaller datasets like ImageNet-21K-P, 2-hops and 3-hops, demonstrating its generalization ability. Our benchmark and code are available at https://kaiyi.me/p/hgrnet.html.