Classifying and Segmenting Microscopy Images Using Convolutional Multiple Instance Learning

TL;DR

This paper introduces a novel CNN-based multiple instance learning approach with a new pooling function, enabling effective classification and segmentation of microscopy images at full resolution without detailed annotations.

Contribution

It presents the adaptive Noisy-AND MIL pooling function and demonstrates end-to-end training of MIL CNNs for microscopy image analysis, outperforming previous methods.

Findings

Outperforms previous methods on microscopy datasets

Robust to outliers with the new pooling function

Enables training with global labels only

Abstract

Convolutional neural networks (CNN) have achieved state of the art performance on both classification and segmentation tasks. Applying CNNs to microscopy images is challenging due to the lack of datasets labeled at the single cell level. We extend the application of CNNs to microscopy image classification and segmentation using multiple instance learning (MIL). We present the adaptive Noisy-AND MIL pooling function, a new MIL operator that is robust to outliers. Combining CNNs with MIL enables training CNNs using full resolution microscopy images with global labels. We base our approach on the similarity between the aggregation function used in MIL and pooling layers used in CNNs. We show that training MIL CNNs end-to-end outperforms several previous methods on both mammalian and yeast microscopy images without requiring any segmentation steps.