Related Experiment Video
Updated: May 5, 2026
07:25
Ovine Lumbar Intervertebral Disc Degeneration Model Utilizing a Lateral Retroperitoneal Drill Bit Injury
Published on: May 25, 2017
11.6K
Exploring deep learning strategies for intervertebral disc herniation detection on veterinary MRI.
Shoujin Huang1, Guoxiong Deng1, Yan Kang1
1Shenzhen Technology University, Shenzhen, China.
Scientific Reports
|July 19, 2024
Summary
This study automates canine intervertebral disc herniation (IVDH) detection in MRI scans using AI. Two-stage models outperformed one-stage detectors, achieving 75.32% average precision for IVDH localization.
Area of Science:
- Veterinary Radiology
- Artificial Intelligence in Veterinary Medicine
- Medical Image Analysis
Background:
- Intervertebral Disc Herniation (IVDH) is a prevalent spinal condition in dogs, affecting their health and mobility.
- Accurate and timely diagnosis of IVDH is crucial for effective treatment and improved patient outcomes.
- Current diagnostic methods for IVDH can be labor-intensive and require specialized expertise.
Purpose of the Study:
- To develop and evaluate artificial intelligence (AI) models for automated detection and localization of IVDH lesions in canine MRI scans.
- To compare the performance of traditional two-stage object detection models against one-stage models for IVDH identification.
- To introduce and assess a novel spinal localization module to enhance IVDH detection accuracy.
Main Methods:
- A dataset of T2-weighted sagittal MRI images from 213 dogs was curated for training and testing AI models.
- Various object detection models, including You Only Look Once X (YOLOX), were implemented and evaluated.
- A novel spinal localization module was developed and integrated into existing detection frameworks.
- Transfer learning techniques were explored to adapt the canine IVDH model for feline imaging.
Main Results:
- Traditional two-stage detection models demonstrated superior performance compared to one-stage models for IVDH detection.
- The integration of the novel spinal localization module improved IVDH detection, achieving an average precision (AP) of up to 75.32%.
- The study identified key challenges and potential strategies for advancing AI in veterinary radiology.
Conclusions:
- AI-powered automated detection and localization of IVDH in canine MRI scans is feasible and promising.
- Two-stage object detection models with a spinal localization module offer high accuracy for veterinary spinal imaging.
- Further research and development are warranted to refine AI tools for veterinary diagnostic imaging.
