microbeSEG dataset
DOI10.5281/zenodo.6497715Zenodo6497715MaRDI QIDQ6703025FDOQ6703025
Dataset published at Zenodo repository.
Bastian Wollenhaupt, Oliver Neumann, Katharina Nöh, Dietrich Kohlheyer, Hanno Scharr, Tim Scherr, Marcel P. Schilling, Ralf Mikut, Johannes Seiffarth
Publication date: 28 April 2022
Copyright license: Creative Commons Attribution 4.0 International
B. subtilis and E. coli cell segmentation dataset consisting of test data annotated by three experts (test), data annotated manually by a single microbeSEG user within 30 minutes (30min-man), data annotated manually by a single microbeSEG user within 30 minutes and data annotated with microbeSEG pre-labeling with 15 minutes manual correction time (30min-man_15min-pre, includes the 30min-man dataset). Images, instance segmentation masks and image-segmentation overlays are provided. All images are crops of size 320px x 320px. Annotations were made with ObiWan-Microbi. Data acquisition The phase contrast images of growing B. subtilis and E. coli colonies were acquired with a fully automated time-lapse microscope setup (TI Eclipse, Nikon, Germany) using a 100x oil immersion objective (Plan Apochromat Oil, N.A. 1.45, WD 170 m, Nikon Microscopy). Time-lapse images were taken every 15 minutes for B. subtilis and every 20 minutes for E. coli. Cultivation took place inside a special microfluidic cultivation device. Resolution: 0.07m/px for B. subtilis und 0.09m/px for E. coli. microbeSEG import For the use with microbeSEG, create or select a new training set within the software and use the training data import functionality. Best import train data with the train checkbox checked, validation data with the val checkbox checked, and test data with the test checkbox checked. Since the images are already normalized, the keep normalization functionality can be used.
This page was built for dataset: microbeSEG dataset
