CytoImageNet: A large-scale pretraining dataset for bioimage transfer learning

TL;DR

CytoImageNet is a large-scale microscopy image dataset designed for pretraining models to improve bioimage analysis, demonstrating competitive performance and capturing unique biological information beyond standard image datasets.

Contribution

The paper introduces CytoImageNet, a new large-scale microscopy dataset for pretraining, which enhances bioimage transfer learning and captures unique biological features.

Findings

Pretraining on CytoImageNet improves microscopy classification performance.

CytoImageNet features contain biological information not present in ImageNet.

The dataset is publicly available for research use.

Abstract

Motivation: In recent years, image-based biological assays have steadily become high-throughput, sparking a need for fast automated methods to extract biologically-meaningful information from hundreds of thousands of images. Taking inspiration from the success of ImageNet, we curate CytoImageNet, a large-scale dataset of openly-sourced and weakly-labeled microscopy images (890K images, 894 classes). Pretraining on CytoImageNet yields features that are competitive to ImageNet features on downstream microscopy classification tasks. We show evidence that CytoImageNet features capture information not available in ImageNet-trained features. The dataset is made available at https://www.kaggle.com/stanleyhua/cytoimagenet.