Feature-Balanced Loss for Long-Tailed Visual Recognition

Mengke Li, Yiu-ming Cheung*, Juyong Jiang

*Corresponding author for this work

Research output: Chapter in book/report/conference proceedingConference proceedingpeer-review

9 Citations (Scopus)


Deep neural networks frequently suffer from performance degradation when the training data is long-tailed because several majority classes dominate the training, resulting in a biased model. Recent studies have made a great effort in solving this issue by obtaining good representations from data space, but few of them pay attention to the influence of feature norm on the predicted results. In this paper, we therefore address the long-tailed problem from feature space and thereby propose the feature-balanced loss. Specifically, we encourage larger feature norms of tail classes by giving them relatively stronger stimuli. Moreover, the stimuli intensity is gradually increased in the way of curriculum learning, which improves the generalization of the tail classes, meanwhile maintaining the performance of the head classes. Extensive experiments on multiple popular long-tailed recognition benchmarks demonstrate that the feature-balanced loss achieves superior performance gains compared with the state-of-the-art methods.
Original languageEnglish
Title of host publication2022 IEEE International Conference on Multimedia and Expo (ICME)
Number of pages6
ISBN (Electronic)9781665485630
ISBN (Print)9781665485647
Publication statusPublished - 18 Jul 2022
Event2022 IEEE International Conference on Multimedia and Expo, ICME 2022 - Taipei, Taiwan
Duration: 18 Jul 202222 Jul 2022

Publication series

NameIEEE International Conference on Multimedia and Expo (ICME)
ISSN (Print)1945-7871
ISSN (Electronic)1945-788X


Conference2022 IEEE International Conference on Multimedia and Expo, ICME 2022
CityTaipei, Taiwan
Internet address

User-Defined Keywords

  • Long-tailed recognition
  • class imbalance learning
  • feature-balanced loss
  • deep neural networks


Dive into the research topics of 'Feature-Balanced Loss for Long-Tailed Visual Recognition'. Together they form a unique fingerprint.

Cite this