Color-$S^{4}L$: Self-supervised Semi-supervised Learning with Image Colorization

TL;DR

This paper introduces Color-$S^{4}L$, a semi-supervised learning framework that leverages image colorization as a self-supervised proxy task, demonstrating improved performance on standard image classification datasets.

Contribution

It proposes a novel semi-supervised learning approach using image colorization as a self-supervised task, with extensive evaluation across multiple network architectures and datasets.

Findings

Effective on CIFAR-10, SVHN, CIFAR-100 datasets

Outperforms previous supervised and semi-supervised methods

Deep evaluation of network architectures

Abstract

This work addresses the problem of semi-supervised image classification tasks with the integration of several effective self-supervised pretext tasks. Different from widely-used consistency regularization within semi-supervised learning, we explored a novel self-supervised semi-supervised learning framework (Color-$S^{4}L$) especially with image colorization proxy task and deeply evaluate performances of various network architectures in such special pipeline. Also, we demonstrated its effectiveness and optimal performance on CIFAR-10, SVHN and CIFAR-100 datasets in comparison to previous supervised and semi-supervised optimal methods.