The Majority Can Help The Minority: Context-rich Minority Oversampling for Long-tailed Classification

TL;DR

This paper introduces a simple yet effective minority oversampling technique that leverages majority class backgrounds to diversify minority samples, significantly improving long-tailed classification performance.

Contribution

The paper proposes a novel context-rich minority oversampling method that enhances minority class diversity by pasting images onto majority class backgrounds, compatible with existing long-tailed recognition methods.

Findings

Achieves state-of-the-art results on long-tailed benchmarks.

Effective without architectural changes or complex algorithms.

Proven through extensive experiments and ablation studies.

Abstract

The problem of class imbalanced data is that the generalization performance of the classifier deteriorates due to the lack of data from minority classes. In this paper, we propose a novel minority over-sampling method to augment diversified minority samples by leveraging the rich context of the majority classes as background images. To diversify the minority samples, our key idea is to paste an image from a minority class onto rich-context images from a majority class, using them as background images. Our method is simple and can be easily combined with the existing long-tailed recognition methods. We empirically prove the effectiveness of the proposed oversampling method through extensive experiments and ablation studies. Without any architectural changes or complex algorithms, our method achieves state-of-the-art performance on various long-tailed classification benchmarks. Our code is made available at https://github.com/naver-ai/cmo.