AI For Glaucoma
This technology has been tested on nearly 60 000 patients’ images at research institutions and hospitals worldwide and has formed the basis for over 20 peer-reviewed publications. It has received consistently strong testimonials from prominent clinicians and KOLs. So far, the deep learning algorithms have been trained and tested on nearly 2000 normal and pathologic OCT scans for glaucoma diagnosis.
The next step is to develop a professional software product for the clinical market. As the software is easily integrated with all OCT and ultrasound devices there are no technical barriers to overcome in developing the software further. The regulatory route is well-defined and straight-forward and the potential market is large as most ophthalmologists routinely use OCT in their practice. There are currently 60,000 OCT machines and 160,000 ultrasound machines on the market. A subscription based revenue model will generate ongoing revenue streams.
It is a standalone image processing software for existing OCT and ultrasound images. It encompasses:
• Novel image enhancer
• Digital staining of tissues
• Artificial intelligence
Stage of Development
TRL 3. Analytical and experimental critical function and characteristics proof-of-concept has been tested on patients’ images.
Initially targeting ophthalmology market, to assist diagnosis of glaucoma. Additional markets can be pursued wherever OCT or ultrasound imaging is used:
• Cardiovascular imaging
• Dermatologic imaging
• Oncologic imaging
• Imaging of internal organs
• Musculoskeletal imaging
• Imaging of developing biologic specimens
Delayed diagnosis and mismanagement of glaucoma can cause vision loss. At present, the detection of glaucomatous structural damage and change is challenging and subjective. NUS researchers have developed software that enhances optical coherence tomography (OCT) images, applies digital staining to tissues and utilises deep learning technologies to improve and simplify assessment of glaucoma.
Our image processing software allows clinicians to visualize deep tissue structures without any surgical interventions and provides highly specific and sensitive diagnosis. The intensity difference between different tissues captured in the OCT image are enhanced to improve definition and each tissue is digitally stained to aid visualization. This allows clinicians to quickly and precisely differentiate tissues and identify abnormalities. Deep learning algorithms are trained to distinguish the key pathologies observed in glaucoma images and provide an aid to clinicians making a diagnosis.
Patent Available. Available for investment into spin out (Abyss Processing Pte Ltd).
Get in Touch