Frozen Feature Augmentation for Few-Shot Image Classification

TL;DR

This paper introduces Frozen Feature Augmentation (FroFA), a simple yet effective method to improve few-shot image classification by applying data augmentations directly in the feature space of pretrained models.

Contribution

The paper presents FroFA, a novel approach to enhance frozen feature-based few-shot learning by applying data augmentations in the feature space, demonstrating consistent improvements.

Findings

FroFA improves few-shot performance across multiple architectures.

Brightness augmentation consistently boosts accuracy.

FroFA is effective across various datasets and pretraining sources.

Abstract

Training a linear classifier or lightweight model on top of pretrained vision model outputs, so-called 'frozen features', leads to impressive performance on a number of downstream few-shot tasks. Currently, frozen features are not modified during training. On the other hand, when networks are trained directly on images, data augmentation is a standard recipe that improves performance with no substantial overhead. In this paper, we conduct an extensive pilot study on few-shot image classification that explores applying data augmentations in the frozen feature space, dubbed 'frozen feature augmentation (FroFA)', covering twenty augmentations in total. Our study demonstrates that adopting a deceptively simple pointwise FroFA, such as brightness, can improve few-shot performance consistently across three network architectures, three large pretraining datasets, and eight transfer datasets.