Related Experiment Video
Updated: May 28, 2025
07:23
Innervation of Human Intestinal Organoids
Published on: January 17, 2025
374
Deliod a lightweight detection model for intestinal organoids based on deep learning.
Yu Sun1, Hanwen Zhang2, Fengliang Huang1
1College of Electrical and Automation Engineering, Nanjing Normal University, Nanjing, 210023, P. R. China.
Scientific Reports
|February 11, 2025
Summary
Deliod, a new YOLOv8-based model, accurately detects intestinal organoid morphology. This streamlined deep learning approach overcomes limitations in current methods for analyzing these crucial biological models.
Area of Science:
- Biotechnology
- Medical Imaging
- Computational Biology
Background:
- Intestinal organoids are vital for studying intestinal disorders.
- Deep learning aids in their morphological analysis.
- Existing methods struggle with overlapping structures and small targets, limiting accuracy.
Purpose of the Study:
- To develop a streamlined deep learning model for automated intestinal organoid detection.
- To improve the accuracy and applicability of organoid morphology analysis.
Main Methods:
- Utilized YOLOv8 architecture to create the Deliod model.
- Applied Deliod to an intestinal organoid dataset for morphological identification.
- Conducted ablation experiments to validate module efficacy.
Main Results:
- Deliod achieved a high mAP50 of 87.5%, outperforming leading detection models.
- Demonstrated improved detection performance through ablation studies.
- The model has a low parameter count (5.41M) and computational load (16.6 GFLOPs).
Conclusions:
- Deliod offers efficient and accurate recognition of intestinal organoid morphology.
- Its streamlined design and low computational requirements facilitate wider application in organoid image analysis.
- The model enhances the utility of organoids in biomedical research by improving analytical capabilities.
