The Majority Can Help The Minority: Context-rich Minority Oversampling for Long-tailed Classification
This addresses the problem of poor generalization in classifiers due to imbalanced data for researchers and practitioners in computer vision, though it is incremental as it builds on existing long-tailed recognition methods.
The paper tackles class imbalance in long-tailed classification by proposing a novel minority oversampling method that pastes minority class images onto rich-context majority class backgrounds to diversify samples, achieving state-of-the-art performance on various benchmarks without architectural changes.
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.