Related Experiment Video
Updated: Jun 30, 2026
09:03
Morphometric Analyses of Shape: The Analysis Software Toolbox for Craniofacial Shape Quantification in Zebrafish
Published on: February 27, 2026
Deep-ZOMA: A Deep Learning-Based Approach for Automated Morphometric Analysis of Zebrafish Larvae Ocular Structures
Youyuan Zhuang1, Chuang Xu1,2, Wei Dai1,2
1State Key Laboratory of Eye Health, Eye Hospital, Wenzhou Medical University, Wenzhou, China.
Translational Vision Science & Technology
|June 29, 2026
Summary
A new deep learning tool, Deep-ZOMA, automates zebrafish ocular measurements, offering faster and more accurate results than manual methods. This innovation aids ocular disease research and drug discovery.
Area of Science:
- Ophthalmology
- Biomedical Engineering
- Computational Biology
Background:
- Zebrafish (Danio rerio) are crucial models in ocular disease research and drug discovery.
- Traditional manual ocular measurements are time-consuming, prone to errors, and lack efficiency.
- There is a need for automated, precise, and high-throughput methods for zebrafish ocular morphometry.
Purpose of the Study:
- To introduce Deep-ZOMA, a deep learning-based tool for automated quantitative measurement of zebrafish larvae ocular structures.
- To validate the accuracy and efficiency of Deep-ZOMA against manual expert measurements.
- To demonstrate the utility of Deep-ZOMA in genetic and drug screening studies.
Main Methods:
- A dual-center dataset of 1820 bright-field zebrafish ocular images was utilized.
- A UNet++ segmentation network was trained using augmentation and a hybrid loss function.
- Sixteen morphometric parameters were computed, and performance was evaluated using Dice coefficient, IoU, and correlation analyses.
Main Results:
- Deep-ZOMA achieved high performance with mean Dice coefficients of 0.96 (internal) and 0.95 (external) and IoU >0.90.
- Automated measurements demonstrated strong correlation and excellent agreement with expert annotations.
- The tool accurately identified ocular phenotypes in a genetic model and was over 20-fold faster than manual measurements.
Conclusions:
- Deep-ZOMA offers a reliable and efficient solution for high-throughput zebrafish ocular morphometry.
- The tool supports applications in ocular genetics, drug screening, and phenotypic studies.
- Deep-ZOMA accelerates translational research by enabling accurate, reproducible ocular outcome assessments relevant to human eye diseases.
