Abstract:This work explores visual recognition models on real-world datasets exhibiting a long-tailed distribution. Most of previous works are based on a holistic perspective that the overall gradient for training model is directly obtained by considering all classes jointly. However, due to the extreme data imbalance in long-tailed datasets, joint consideration of different classes tends to induce the gradient distortion problem; i.e., the overall gradient tends to suffer from shifted direction toward data-rich classes and enlarged variances caused by data-poor classes. The gradient distortion problem impairs the training of our models. To avoid such drawbacks, we propose to disentangle the overall gradient and aim to consider the gradient on data-rich classes and that on data-poor classes separately. We tackle the long-tailed visual recognition problem via a dual-phase-based method. In the first phase, only data-rich classes are concerned to update model parameters, where only separated gradient on data-rich classes is used. In the second phase, the rest data-poor classes are involved to learn a complete classifier for all classes. More importantly, to ensure the smooth transition from phase I to phase II, we propose an exemplar bank and a memory-retentive loss. In general, the exemplar bank reserves a few representative examples from data-rich classes. It is used to maintain the information of data-rich classes when transiting. The memory-retentive loss constrains the change of model parameters from phase I to phase II based on the exemplar bank and data-poor classes. The extensive experimental results on four commonly used long-tailed benchmarks, including CIFAR100-LT, Places-LT, ImageNet-LT, and iNaturalist 2018, highlight the excellent performance of our proposed method.

Classification of Long-Tailed Data Based on Bilateral-Branch Generative Network with Time-Supervised Strategy

Cumulative Dual-Branch Network Framework for Long-Tailed Multi-Class Classification

BBN: Bilateral-Branch Network with Cumulative Learning for Long-Tailed Visual Recognition

SAU: A Dual-Branch Network to Enhance Long-Tailed Recognition via Generative Models

A dual-branch model with inter- and intra-branch contrastive loss for long-tailed recognition

Improving Long-Tailed Classification from Instance Level

A Dual-Branch Learning Model with Gradient-Balanced Loss for Long-Tailed Multi-Label Text Classification

On Size-Oriented Long-Tailed Graph Classification of Graph Neural Networks

A Dual Progressive Strategy for Long-Tailed Visual Recognition

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

Long-Tailed Recognition of Food Crop Disease Images Based on Deep Neural Networks: Long-Tailed Recognition of Food Crop Disease Images.

Relieving the Incompatibility of Network Representation and Classification for Long-Tailed Data Distribution

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

Long-tailed Visual Recognition via Gaussian Clouded Logit Adjustment

Dual-branch Branch Networks Based on Contrastive Learning for Long-Tailed Remote Sensing

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

Contrastive Learning Based Hybrid Networks for Long-Tailed Image Classification

Tail Classes Matter: Long-Tailed Object Detection Revisited.

LCReg: Long-Tailed Image Classification with Latent Categories based Recognition

Text-Guided Diverse Image Synthesis for Long-Tailed Remote Sensing Object Classification

Bt-Vmf Contrastive and Collaborative Learning for Long-Tailed Visual Recognition