SCAN: Learning to Classify Images without Labels

TL;DR

This paper introduces SCAN, a two-step unsupervised image classification method that uses self-supervised features and learnable clustering, significantly outperforming previous approaches and working effectively on large-scale datasets like ImageNet.

Contribution

SCAN proposes a decoupled two-step approach for unsupervised image classification, combining self-supervised feature learning with a learnable clustering method, improving accuracy and scalability.

Findings

Outperforms state-of-the-art methods by large margins on CIFAR and STL10.

Achieves promising results on ImageNet without ground-truth labels.

Outperforms semi-supervised methods in low-data regimes.

Abstract

Can we automatically group images into semantically meaningful clusters when ground-truth annotations are absent? The task of unsupervised image classification remains an important, and open challenge in computer vision. Several recent approaches have tried to tackle this problem in an end-to-end fashion. In this paper, we deviate from recent works, and advocate a two-step approach where feature learning and clustering are decoupled. First, a self-supervised task from representation learning is employed to obtain semantically meaningful features. Second, we use the obtained features as a prior in a learnable clustering approach. In doing so, we remove the ability for cluster learning to depend on low-level features, which is present in current end-to-end learning approaches. Experimental evaluation shows that we outperform state-of-the-art methods by large margins, in particular +26.6% on CIFAR10, +25.0% on CIFAR100-20 and +21.3% on STL10 in terms of classification accuracy. Furthermore, our method is the first to perform well on a large-scale dataset for image classification. In particular, we obtain promising results on ImageNet, and outperform several semi-supervised learning methods in the low-data regime without the use of any ground-truth annotations. The code is made publicly available at https://github.com/wvangansbeke/Unsupervised-Classification.