Related Experiment Video
Updated: Aug 5, 2026
08:56
Automatic Image Processing to Determine the Community Size Structure of Riverine Macroinvertebrates
Published on: January 13, 2023
ReViTA-Unet: An Enhanced Semantic Segmentation Model for Automated Morphometric Analysis of Macrobrachium rosenbergii
Dawei Sun1,2, Qi Chen1,3, Guanghui Yu1,2
1Institute of Agricultural Equipment, Zhejiang Academy of Agricultural Sciences, Hangzhou 310021, China.
Sensors (Basel, Switzerland)
|July 28, 2026
Summary
This study introduces ReViTA-UNet, an automated system for prawn morphometrics. It accurately measures traits, improving precision aquaculture and selective breeding efficiency.
Area of Science:
- Aquaculture
- Computer Vision
- Biotechnology
Background:
- Manual morphometric analysis of *Macrobrachium rosenbergii* (prawns) is inefficient and operator-dependent.
- Accurate measurements are crucial for selective breeding, growth monitoring, and precision aquaculture.
- Existing methods lack automation and accuracy for complex biological structures.
Purpose of the Study:
- To develop an automated, non-contact framework for accurate morphometric analysis of *Macrobrachium rosenbergii*.
- To enhance semantic segmentation for precise extraction of prawn morphological traits.
- To establish an efficient system for automated prawn phenotyping.
Main Methods:
- Developed ReViTA-UNet, an enhanced semantic segmentation network.
- Integrated complementary feature extraction for improved segmentation of elongated structures and complex boundaries.
- Created an automated measurement system to convert segmented images into morphometric parameters.
Main Results:
- ReViTA-UNet achieved high segmentation accuracy (Dice: 97.7%, mIoU: 96.7%).
- Outperformed eight other semantic segmentation models in accuracy.
- Automated system showed high agreement with manual measurements (R² = 0.987) with low error (1.83% for body length).
Conclusions:
- ReViTA-UNet offers an accurate and efficient solution for automated prawn phenotyping.
- The framework supports precision aquaculture and selective breeding initiatives.
- Provides a foundation for intelligent aquaculture applications, pending commercial environment validation.
