Related Experiment Video
Updated: Jun 18, 2026
07:21
Automated, High-Throughput Detection of Bacterial Adherence to Host Cells
Published on: September 17, 2021
3.5K
Annotated dataset for deep-learning-based bacterial colony detection.
László Makrai1, Bettina Fodróczy1,2, Sára Ágnes Nagy2
1Department of Microbiology and Infectious Diseases, University of Veterinary Medicine, 1143, Budapest, Hungary.
Scientific Data
|July 28, 2023
Summary
Automating bacterial colony counting using artificial intelligence significantly speeds up a common microbiology task. This study provides a valuable dataset for developing AI tools to count bacterial colonies, aiding research in veterinary importance and food hygiene.
Area of Science:
- Microbiology
- Artificial Intelligence
- Computer Vision
Background:
- Accurate bacterial quantification is crucial in fields like infectiology and food hygiene.
- Traditional bacterial colony counting from cultures is manual, time-consuming, and labor-intensive.
- Automating this process can improve efficiency and accuracy in microbiological studies.
Purpose of the Study:
- To develop and evaluate an automated method for counting bacterial colonies using convolutional neural networks (CNNs).
- To create a comprehensive dataset of bacterial cultures for training and validating AI models.
Main Methods:
- Cultured 24 bacterial species of veterinary importance at various concentrations on solid media.
- Captured 369 digital images of the bacterial cultures.
- Manually annotated a total of 56,865 bacterial colonies using bounding boxes.
Main Results:
- A large, annotated dataset of bacterial colonies was generated.
- The dataset comprises images of diverse bacterial species and concentrations.
- This dataset is suitable for training AI models for automated colony counting.
Conclusions:
- The developed dataset is a valuable resource for advancing AI-driven automation in bacterial colony counting.
- This work facilitates the development of more efficient and accurate microbiological analysis tools.
- The dataset will support future research in artificial intelligence applications for microbiology.
