A Fundus-Based Color Mapping Method to Optimize RNFL Clock Hours for Glaucoma Screening.

Area of Science:

• Ophthalmology
• Medical Imaging
• Computer Vision

Background:

• Glaucoma is a primary cause of irreversible blindness, underscoring the need for early detection and monitoring.
• Accurate representation of the retinal nerve fiber layer (RNFL) is crucial for diagnosing glaucoma.
• Previous methods for RNFL visualization had limitations in depicting critical thin regions.

Purpose of the Study:

• To develop and evaluate a novel color mapping technique and a new loss function for improved RNFL clock hour representation.
• To enhance the visualization of RNFL thickness, particularly in areas critical for glaucoma screening.
• To assess the performance of the proposed method using the Chula-Thammasat Glaucoma dataset (GlauCUTU-DATA).

Main Methods:

• A novel color mapping technique was developed, assigning colors to RNFL thickness ranges.
• A detail-gained color map was employed to improve visualization of thinner RNFL regions.
• Peak Signal-to-Noise Ratio (PSNR) was used as the loss function, replacing Root Mean Square Error (RMSE), for more detailed image quality evaluation.

Main Results:

• The novel color mapping and PSNR loss function demonstrated effectiveness in generating RNFL clock hours.
• An overall micro F1-score of 70.51% was achieved across all clock positions.
• Specific clock positions (2, 5, 7, and 12), critical for glaucoma screening, achieved micro-average F1-scores exceeding 80%.

Conclusions:

• The developed color mapping methodology and PSNR loss function show promise for advancing glaucoma diagnosis and management.
• The technique offers improved visualization and accurate representation of RNFL thickness, especially in diagnostically critical areas.
• This approach represents a potential step forward in the early detection and monitoring of glaucoma.