Automatic image annotation for fluorescent cell nuclei segmentation.

Iris E Ruider,Andreas R Bausch,Fabian Englbrecht

doi:10.1371/journal.pone.0250093

Iris E Ruider, Andreas R Bausch + Show 1 more

Open Access

https://doi.org/10.1371/journal.pone.0250093

Copy DOI

Abstract

Dataset annotation is a time and labor-intensive task and an integral requirement for training and testing deep learning models. The segmentation of images in life science microscopy requires annotated image datasets for object detection tasks such as instance segmentation. Although the amount of annotated image data has been steadily reduced due to methods such as data augmentation, the process of manual or semi-automated data annotation is the most labor and cost intensive task in the process of cell nuclei segmentation with deep neural networks. In this work we propose a system to fully automate the annotation process of a custom fluorescent cell nuclei image dataset. By that we are able to reduce nuclei labelling time by up to 99.5%. The output of our system provides high quality training data for machine learning applications to identify the position of cell nuclei in microscopy images. Our experiments have shown that the automatically annotated dataset provides coequal segmentation performance compared to manual data annotation. In addition, we show that our system enables a single workflow from raw data input to desired nuclei segmentation and tracking results without relying on pre-trained models or third-party training datasets for neural networks.

Highlights

Objection detection becomes of increasing importance in diverse fields of life science microscopy, requiring segmenting, counting or tracking individual cell nuclei
We investigate the performance of a manually annotated microscopy image training dataset compared to our automated data annotation
Our method provides a solution to automate the data annotation process of cell nuclei in microscopy images

Summary

Introduction

Objection detection becomes of increasing importance in diverse fields of life science microscopy, requiring segmenting, counting or tracking individual cell nuclei. With the current deep learning models individual objects of interest such as nuclei of cells are detectable on a pixel level Thereby, deep learning methods perform better than traditional cell nuclei detection and segmentation algorithms such as watershed-segmentation [5,6]. Segmentation using supervised machine learning techniques requires two steps: data annotation and training. The labelled data is used together with the associated images as input data to train a machine learning model [5]. Data annotation is often still done manually or can sometimes be semi-automated depending on the morphological complexity of the object and requiring manual fine-

Methods

Results

Discussion

Conclusion