Abstract:Conventional detectors suffer from performance degradation when dealing with long-tailed data due to a classification bias towards the majority head categories. In this paper, we contend that the learning bias originates from two factors: 1) the unequal competition arising from the imbalanced distribution of foreground categories, and 2) the lack of sample diversity in tail categories. To tackle these issues, we introduce a unified framework called BAlanced CLassification (BACL), which enables adaptive rectification of inequalities caused by disparities in category distribution and dynamic intensification of sample diversities in a synchronized manner. Specifically, a novel foreground classification balance loss (FCBL) is developed to ameliorate the domination of head categories and shift attention to difficult-to-differentiate categories by introducing pairwise class-aware margins and auto-adjusted weight terms, respectively. This loss prevents the over-suppression of tail categories in the context of unequal competition. Moreover, we propose a dynamic feature hallucination module (FHM), which enhances the representation of tail categories in the feature space by synthesizing hallucinated samples to introduce additional data variances. In this divide-and-conquer approach, BACL sets a new state-of-the-art on the challenging LVIS benchmark with a decoupled training pipeline, surpassing vanilla Faster R-CNN with ResNet-50-FPN by 5.8% AP and 16.1% AP for overall and tail categories. Extensive experiments demonstrate that BACL consistently achieves performance improvements across various datasets with different backbones and architectures. Code and models are available at <a class="link-external link-https" href="https://github.com/Tianhao-Qi/BACL" rel="external noopener nofollow">this https URL</a>.

Learning Box Regression and Mask Segmentation under Long-Tailed Distribution with Gradient Transfusing

Adaptive Class Suppression Loss for Long-Tail Object Detection

The Equalization Losses: Gradient-Driven Training for Long-tailed Object Recognition

Boosting Long-tailed Object Detection via Step-wise Learning on Smooth-tail Data

Class-Guided Triple Head Prediction Network for Long-Tail Object Detection

Long-tailed Visual Recognition via Gaussian Clouded Logit Adjustment

Rectify the Regression Bias in Long-Tailed Object Detection

Balanced Classification: A Unified Framework for Long-Tailed Object Detection

LTRL: Boosting Long-tail Recognition via Reflective Learning

Divide and Retain: A Dual-Phase Modeling for Long-Tailed Visual Recognition

Calibrating Class Activation Maps for Long-Tailed Visual Recognition

Long-Tailed Object Detection Pre-training: Dynamic Rebalancing Contrastive Learning with Dual Reconstruction

Gradient-Aware Logit Adjustment Loss for Long-tailed Classifier

A Dual Progressive Strategy for Long-Tailed Visual Recognition

GREB: Gradient Re-Balanced Loss for Long-Tailed Multi-Lable Classification

Text-guided Fourier Augmentation for long-tailed recognition

Asynchronous Modeling: A Dual-phase Perspective for Long-Tailed Recognition

Transformer Fusion and Residual Learning Group Classifier Loss for Long-Tailed Traffic Sign Detection

Text-Guided Mixup Towards Long-Tailed Image Categorization

Margin-aware Rectified Augmentation for Long-Tailed Recognition

Long-tailed Visual Recognition with Deep Models: A Methodological Survey and Evaluation