portability

AI-Based AMD DETECTIon from

 PortabLE OCT IMAGES

Portable OCT devices are much cheaper and easier to operate, but suffers from low SNR, small field of view and limited resolution.

High-quality commercial optical coherence tomography (OCT) systems can cost over $50K and are bulky, weighing over 50 pounds.

We have developed an image processing pipeline to super-resolve images captured from a low-cost, portable Optical Coherence Tomography device, enabling downstream diagnosis such as Age-related Macular Degeneration detection.  

The OptiRes-DDPM is based on Denoising Diffusion Probabilistic Models, which achieves  state-of-the-art performance in  generative modeling. 


It addresses limitations seen in previous methods such as Variational Autoencoders (VAEs), Generative Adversarial Networks (GANs), and autoregressive models. 




Our invention ultimately is aimed to be integrated as software into commercial portable OCT devices, enhancing its image resolution and quality in real-time. With this pioneering approach in the field of retinal imaging and its subsequent deployment in clinics, our computational tool has the potential to revolutionize the current workflow of ophthalmic diagnostics.