Distribution Alignment: A Unified Framework for Long-tail Visual Recognition

TL;DR

This paper introduces a unified distribution alignment framework that adaptively calibrates classification scores and balances class priors, significantly improving long-tail visual recognition across multiple tasks.

Contribution

It proposes a novel adaptive calibration and re-weighting strategy that unifies long-tail recognition methods for diverse visual tasks.

Findings

Achieves state-of-the-art results on image classification, semantic segmentation, object detection, and instance segmentation.

Demonstrates the effectiveness of the unified framework across multiple long-tail recognition scenarios.

Provides a simple yet powerful approach with extensive experimental validation.

Abstract

Despite the recent success of deep neural networks, it remains challenging to effectively model the long-tail class distribution in visual recognition tasks. To address this problem, we first investigate the performance bottleneck of the two-stage learning framework via ablative study. Motivated by our discovery, we propose a unified distribution alignment strategy for long-tail visual recognition. Specifically, we develop an adaptive calibration function that enables us to adjust the classification scores for each data point. We then introduce a generalized re-weight method in the two-stage learning to balance the class prior, which provides a flexible and unified solution to diverse scenarios in visual recognition tasks. We validate our method by extensive experiments on four tasks, including image classification, semantic segmentation, object detection, and instance segmentation. Our approach achieves the state-of-the-art results across all four recognition tasks with a simple and unified framework. The code and models will be made publicly available at: https://github.com/Megvii-BaseDetection/DisAlign