MUMBAI, India, Aug. 31 -- Intellectual Property India has published a patent application (202611071479 A) filed by Manaktala, Siddhartha; and Manaktala, Payal on June 09, 2026, for Multi-Model Artificial Intelligence System And Method For Ophthalmic Image Classification On Smartphone Devices.
Inventor includes Manaktala, Siddhartha.
The application for the patent was published on August 28, 2026, under issue no. 35/2026.
Abstract: An ophthalmic screening system (100) and method are disclosed, comprising a processor (102), smartphone camera (104), posterior imaging pathway (106), and anterior imaging pathway (108). A first inference module (112) with Inception V3 CNN (114) classifies Diabetic Retinopathy and Glaucoma from fundus images. A cascade execution controller (130) first executes a second inference module (118) with YOLO11n segmentation model (120) generating scleral (122) and iris (124) masks and a red pixel ratio via HSV dual hue-range isolation; the ratio serves as diagnostic indicator and routing gate. Where the ratio is below 0.4, a third inference module (126) with YOLO11n detection model (128) classifies Pterygium, Cataract Clouding, and Stye; otherwise an inflammatory condition is identified. A clinical validation module (132) administers a symptom questionnaire (134). A cloud inference backend (136) with unified API (138) processes images, and separate structured patient-facing reports (146) are delivered via a consumer mobile application (148).
Disclaimer: Curated by HT Syndication.
