New Thailand medical innovation involving an AI-algorithm based automated grading system for corneal ulcers

Thailand Medical: A groundbreaking innovation in Thailand’s medical field is set to revolutionize how corneal ulcers are diagnosed and treated. Researchers from the Faculty of Engineering at Chiang Mai University have developed an AI algorithm-based automated grading system for corneal ulcers. This innovative system utilizes deep learning and the Hough Circle Transform to offer more accurate and efficient diagnoses of this common ocular condition. This Thailand Medical news report explores how this new technology works, its impact on medical practice, and the potential benefits for patients and ophthalmologists alike.


New Thailand medical innovation involving an AI-algorithm based automated
grading system for corneal ulcers

Understanding Corneal Ulcers
Corneal ulcers are a serious eye condition that can result from infection, injury, or inflammation. These ulcers appear as open sores on the cornea, the clear front surface of the eye, and can cause symptoms such as eye pain, redness, and impaired vision. Diagnosing and grading the severity of corneal ulcers is essential for providing appropriate treatment and preventing further complications, such as vision loss.
 
Traditionally, ophthalmologists have relied on manual assessment methods to evaluate corneal ulcers, which involve examining ocular staining images and comparing them with severity indices. This process is not only time-consuming but can also lead to inconsistent results. Different evaluators may offer varying assessments, and even the same evaluator might produce different results when grading the same images at different times.
 
To address these challenges, researchers from Chiang Mai University have developed a new automated system. This AI-powered system has the potential to transform the way ophthalmologists diagnose corneal ulcers by providing more consistent, reliable, and efficient results.
 
The AI Algorithm-Based System
The new system is based on a segmentation-based automated grading model that leverages deep learning techniques. The algorithm operates in two main stages: cornea segmentation and corneal ulcer segmentation.

In the cornea segmentation stage, the system uses a deep learning model in conjunction with the Hough Circle Transform to accurately detect the corneal area from ocular staining images. By doing so, it creates a mask for the corneal area, which is then used in the subsequent stage of the process.
 
In the corneal ulcer segmentation stage, the AI system applies another deep learning model that uses the predicted corneal mask from the first stage as additional training data. This approach allows the system to focus on the relevant areas and accurately segment the ulcerated regions. The system ultimately provides two key outputs: the percentage of the corneal area affected by the ulcer and the severity of the ulcer, graded according to the Type-Grade (TG) grading standard.
 
By automating the grading process, the system eliminates the inconsistencies associated with manual grading and ensures that patients receive a more accurate diagnosis. This articl e highlights the significant advantages that this innovation offers to both patients and ophthalmologists, streamlining the diagnosis process and enabling more timely treatment interventions.
 
Study Findings and Key Results
The research team tested the AI algorithm-based system using the SUSTech-SYSU public dataset, which contains a large collection of ocular staining images. The system demonstrated impressive performance, achieving an Intersection over Union (IoU) of 89.23% for corneal segmentation and 82.94% for corneal ulcer segmentation. Additionally, the system’s ability to estimate the percentage of the corneal area affected by the ulcer had a Mean Absolute Error of just 2.51%. In terms of accuracy for severity grading, the system performed at an impressive rate of 86.15%, demonstrating its potential for real-world clinical applications.
 
These findings show that the AI system can accurately segment corneal ulcers and grade their severity with high precision, outperforming traditional manual methods in both speed and consistency. The study emphasizes how this innovation could reduce the workload of ophthalmologists and improve patient outcomes by enabling more reliable and timely diagnoses.
 
The Impact on Ophthalmology
Ophthalmologists often face a significant challenge when diagnosing corneal ulcers, particularly when it comes to grading the severity of the condition. The manual grading process is subjective and can vary from one specialist to another. Furthermore, repeated assessments of the same images by the same evaluator can yield different results. This inconsistency can affect treatment decisions and, ultimately, patient outcomes.
 
The introduction of this AI algorithm-based system is expected to greatly enhance diagnostic efficiency in ophthalmology. By providing consistent and dependable results, the system minimizes the risk of human error and offers more robust handling of variations in eye size. As a result, ophthalmologists can make more informed decisions and provide patients with more precise treatment plans.
 
Another important aspect of this innovation is its ability to save time. With the AI system automating the grading process, ophthalmologists can focus more on patient care rather than spending valuable time on manual image analysis. This increased efficiency can help streamline clinical workflows, particularly in busy eye care clinics or hospitals.
 
Future Potential and Expanding Applications
While this AI-powered grading system is currently focused on corneal ulcers, its underlying technology has the potential for broader applications in ophthalmology and beyond. The deep learning algorithms used in this system could be adapted to diagnose and grade other ocular conditions, such as cataracts, glaucoma, or diabetic retinopathy. As the system continues to evolve, it could become an essential tool in the diagnosis and treatment of various eye diseases, improving overall patient care.
 
Moreover, the system’s ability to work with large datasets and generate highly accurate results could make it useful in other medical fields that require image analysis and segmentation. For example, similar systems could be developed to assist in diagnosing skin conditions, detecting tumors in radiology images, or even assessing cardiovascular health through imaging.
 
Conclusion
The introduction of this AI algorithm-based automated grading system for corneal ulcers represents a major step forward in Thailand’s medical innovation. Developed by researchers from the Faculty of Engineering at Chiang Mai University, this system offers significant improvements in diagnostic efficiency, consistency, and accuracy for corneal ulcer assessments. By automating the grading process, the system eliminates the inconsistencies associated with manual evaluations and provides ophthalmologists with more reliable data to inform their treatment decisions.
 
This innovation not only benefits ophthalmologists by reducing their workload but also enhances patient care by ensuring timely and precise diagnoses. As the system continues to develop and expand into other areas of medical diagnosis, it has the potential to revolutionize how various conditions are diagnosed and treated.
 
The study findings were published in the peer-reviewed journal: Algorithms.
https://www.mdpi.com/1999-4893/17/9/405
 
For the latest Thailand Medical innovations, keep on logging to Thailand Medical News.
 
Read Also:
https://www.thailandmedical.news/news/thailand-medical-researchers-discover-nutgalls-from-quercus-infectoria-speed-up-wound-healing
 
https://www.thailandmedical.news/news/thailand-medical-researchers-develop-new-nanocarrier-to-deliver-curcumin-directly-to-breast-cancer-cells