Synthetic patches, real images: screening for centrosome aberrations in EM images of human cancer cells

TL;DR

This paper introduces an automated, high-accuracy algorithm for detecting centrioles in electron microscopy images of cancer cells, leveraging synthetic patches and real images to improve detection performance.

Contribution

The authors develop a hybrid training approach combining synthetic patches and real images, significantly enhancing centriole detection accuracy in EM images.

Findings

The method outperforms previous approaches on real patient data.

Synthetic patches improve training effectiveness.

High accuracy detection is achieved with a two-level DenseNet.

Abstract

Recent advances in high-throughput electron microscopy imaging enable detailed study of centrosome aberrations in cancer cells. While the image acquisition in such pipelines is automated, manual detection of centrioles is still necessary to select cells for re-imaging at higher magnification. In this contribution we propose an algorithm which performs this step automatically and with high accuracy. From the image labels produced by human experts and a 3D model of a centriole we construct an additional training set with patch-level labels. A two-level DenseNet is trained on the hybrid training data with synthetic patches and real images, achieving much better results on real patient data than training only at the image-level. The code can be found at https://github.com/kreshuklab/centriole_detection.