AI for Ophthalmological
Measurement System
Challenge
There was a need for a solution that allows precise and accurate assessments of proptosis patients from their homes, reducing the need for pre-surgery hospital visits and providing real-time post-surgery evaluations.
Solution
Glorium Technologies developed an end-to-end ophthalmological measurement system using advanced computer vision segmentation models. Our model outperforms Google Facemesh, providing more accurate and reliable measurements.
Data:
Actual and synthetic images of faces generated by AI models.
Features:
Our developers developed an AI model to generate synthetic images of faces. These images were then auto-labeled and preprocessed to adapt to CV models, ensuring high-quality training data and robust model performance.
CV segmentation model
To guarantee precise and accurate measurements, we implemented advanced segmentation models capable of accurately identifying and measuring various parts of the eye, including the eyelids, eyebrows, iris, and sclera.
Web/smartphone app:
The solution includes a user-friendly web and smartphone app, allowing patients to conduct assessments at home without the need for hospital visits.
Business Value
Improved patient outcomesWeb/smartphone app eliminates the need for pre-surgery hospital visits and ensures that patients can access the tool anytime, anywhere.
Boosting revenueBy streamlining the assessment process and reducing the need for in-person visits, the hospital increased its capacity to serve more patients. This efficiency led to a significant boost in revenue.
Competitive edgeImplementing CV models and AI-generated synthetic images makes the medical company a leader in innovative healthcare solutions.
Tech Stack:
AI
Python
PyTorch
AWS
SQL Server
Scikit Learn
Our clients journeys
