Normalizing Batch Normalization for Long-Tailed Recognition

2025-01-06

IEEE Transactions on Image Processing

AI Summary

To address the poor minority-class recognition performance of deep models under long-tailed distributions, this paper identifies directional–magnitude imbalance in Batch Normalization (BN) layer parameters as the root cause of feature bias. We propose a decoupled BN reconstruction method: after vectorizing BN weights and biases, we apply L₂ normalization to enforce direction uniformity, while introducing learnable scalar multipliers to control magnitude—enabling class-agnostic feature strength equalization under end-to-end supervision. Our approach requires no architectural modifications or loss-function alterations, yet significantly enhances representation learning for rare classes. Extensive experiments demonstrate state-of-the-art performance on CIFAR-10-LT, CIFAR-100-LT, ImageNet-LT, and iNaturalist 2018. The implementation is publicly available.

Technology Category

• Machine Learning: Multi-class/Multi-label Learning & Extreme Classification
• Computer Vision: Representation Learning for Vision
• Natural Language Processing: Learning & Optimization for NLP

Application Category

• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Web Mining and Content Analysis: Normalization, clustering, classification, and summarization of Web text

Abstract

In real-world scenarios, the number of training samples across classes usually subjects to a long-tailed distribution. The conventionally trained network may achieve unexpected inferior performance on the rare class compared to the frequent class. Most previous works attempt to rectify the network bias from the data-level or from the classifier-level. Differently, in this paper, we identify that the bias towards the frequent class may be encoded into features, i.e., the rare-specific features which play a key role in discriminating the rare class are much weaker than the frequent-specific features. Based on such an observation, we introduce a simple yet effective approach, normalizing the parameters of Batch Normalization (BN) layer to explicitly rectify the feature bias. To achieve this end, we represent the Weight/Bias parameters of a BN layer as a vector, normalize it into a unit one and multiply the unit vector by a scalar learnable parameter. Through decoupling the direction and magnitude of parameters in BN layer to learn, the Weight/Bias exhibits a more balanced distribution and thus the strength of features becomes more even. Extensive experiments on various long-tailed recognition benchmarks (i.e., CIFAR-10/100-LT, ImageNet-LT and iNaturalist 2018) show that our method outperforms previous state-of-the-arts remarkably. The code and checkpoints are available at https://github.com/yuxiangbao/NBN.

Problem

• Long-tailed distribution
• Deep learning model
• Minority class recognition

Innovation

• Batch Normalization Adjustment
• Class-balanced Feature Strength
• Long-tailed Dataset Recognition