Li, Zhen;
Liu, Zhongyuan;
Chang, Dongliang;
Sain, Aneeshan;
Li, Xiaoxu;
Ma, Zhanyu;
Xue, Jing-Hao;
(2025)
Self-randomized focuses effectively boost metric-based few-shot classifiers.
Pattern Recognition
, 164
, Article 111538. 10.1016/j.patcog.2025.111538.
![[thumbnail of ZhenLi-ZhongyuanLiu-PR-2025.pdf]](https://discovery.ucl.ac.uk/style/images/fileicons/text.png) |
Text
ZhenLi-ZhongyuanLiu-PR-2025.pdf - Accepted Version Access restricted to UCL open access staff until 12 March 2026. Download (5MB) |
Abstract
Towards solving a few-shot image classification task, deep metric learning is the de-facto approach. Usually the idea here is to train a deep metric model on base data, and evaluate it using novel data without any fine-tuning. Enhancing model performance is mostly focused upon improving feature or class representations, or designing or learning new metrics, often ignoring deep exploration of data-augmentation techniques to enhance few-shot learning. Interestingly, we discover that augmentation strategies, such as Cutout, Mixup and CutMix, would in fact greatly enhance performance of few-shot models. We conjecture, this is because such augmentation techniques encourage the model to extend its focus on multiple discriminative regions of an object instead of restricting to just the single-most discriminative point. Following this important discovery, we propose two simple yet effective novel data augmentation methods, viz. CutRot and CutCov, specifically designed to self-randomize focuses within an image itself for metric-based few-shot image classification. While CutRot involves random rotation of any patch within the image, CutCov focuses on random swapping of patches, again within the image. Extensive experiments verify that CutRot or CutCov can significantly boost performances of both classic and recent popular metric-based methods and performs much better than other augmentation methods of Cutout, Mixup, and CutMix on four few-shot image classification datasets. Code is available at https://github.com/liz-lut/CutRot-and-CutCov-main
| Type: | Article |
|---|---|
| Title: | Self-randomized focuses effectively boost metric-based few-shot classifiers |
| DOI: | 10.1016/j.patcog.2025.111538 |
| Publisher version: | https://doi.org/10.1016/j.patcog.2025.111538 |
| Language: | English |
| Additional information: | This version is the author-accepted manuscript. For information on re-use, please refer to the publisher’s terms and conditions. |
| Keywords: | Few-shot learning, Fine-grained image classification, Data augmentation, Self-randomized focuses |
| UCL classification: | UCL UCL > Provost and Vice Provost Offices > UCL BEAMS UCL > Provost and Vice Provost Offices > UCL BEAMS > Faculty of Maths and Physical Sciences UCL > Provost and Vice Provost Offices > UCL BEAMS > Faculty of Maths and Physical Sciences > Dept of Statistical Science |
| URI: | https://discovery.ucl.ac.uk/id/eprint/10206926 |
Archive Staff Only
 |
View Item |
