Related Experiment Video
Updated: Jun 20, 2025
04:48
Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
Published on: November 30, 2022
2.7K
Automated identification and segmentation of urine spots based on deep-learning
Peerj
|July 22, 2024
Summary
A new deep learning system automates rodent urine spot analysis, overcoming manual limitations. This method precisely identifies and quantifies urination events, improving the study of micturition behavior.
Area of Science:
- Physiology
- Neuroscience
- Computer Science
Background:
- Micturition is vital for waste elimination and homeostasis.
- Traditional urine spot assays (VSA) in rodents are manual, subjective, and struggle with quantitative analysis, especially for overlapping spots.
Purpose of the Study:
- To develop a deep learning-based system for automated identification and segmentation of rodent urine spots.
- To overcome the limitations of traditional VSA, including subjective errors and difficulties in quantifying overlapping spots and urination timing.
Main Methods:
- Utilized a deep learning approach combining target detection and instance segmentation networks.
- Developed a system for automatic identification, precise segmentation, and quantification of urine spots in rodents.
Main Results:
- Achieved automated detection of urine spot area, significantly reducing subjective errors compared to traditional VSA.
- Accurately determined urination time for each spot and effectively quantified overlapping spots.
- Enabled high-throughput and precise urine spot detection.
Conclusions:
- The deep learning system provides a robust, automated solution for analyzing rodent urination behavior.
- Offers crucial technical support for studying the neural mechanisms of micturition.
- Enhances the precision and efficiency of urine spot analysis in research.
