# Class-Difficulty Based Methods for Long-Tailed Visual Recognition

Source: [Class-Difficulty Based Methods for Long-Tailed Visual Recognition](<https://www.sota2.com/research/paper/68fa6d4b1e209131cea38d79>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2025-10-26T00:30:35.914Z

Authors: Saptarshi Sinha, Hiroki Ohashi, Katsuyuki Nakamura

Published: 2022-07-29T06:33:22.000Z

## Abstract

Long-tailed datasets are very frequently encountered in real-world use cases where few classes or categories (known as majority or head classes) have higher number of data samples compared to the other classes (known as minority or tail classes). Training deep neural networks on such datasets gives results biased towards the head classes. So far, researchers have come up with multiple weighted loss and data re-sampling techniques in efforts to reduce the bias. However, most of such techniques assume that the tail classes are always the most difficult classes to learn and therefore need more weightage or attention. Here, we argue that the assumption might not always hold true. Therefore, we propose a novel approach to dynamically measure the instantaneous difficulty of each class during the training phase of the model. Further, we use the difficulty measures of each class to design a novel weighted loss technique called \`class-wise difficulty based weighted (CDB-W) loss' and a novel data sampling technique called \`class-wise difficulty based sampling (CDB-S)'. To verify the wide-scale usability of our CDB methods, we conducted extensive experiments on multiple tasks such as image classification, object detection, instance segmentation and video-action classification. Results verified that CDB-W loss and CDB-S could achieve state-of-the-art results on many class-imbalanced datasets such as ImageNet-LT, LVIS and EGTEA, that resemble real-world use cases.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Image Classification on ImageNet-LT (test)](<https://www.sota2.com/research/sota/image-classification-on-imagenet-lt-test.md>): Image Classification; ImageNet-LT (test); Top-1 Acc (All): not recorded.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2207.14499>)

[PDF](<https://arxiv.org/pdf/2207.14499.pdf>)
