MUMBAI, India, Aug. 17 -- Intellectual Property India has published a patent application (202641095205 A) filed by Amrita Vishwa Vidyapeetham on August 06, 2026, for Dermoscopic Skin-Lesion Classification Using Shared Base Classifiers For Cyber-Attack Detection, Mitigation, And Recovery.
Inventors include Ravichandran, Harini Shree; Kaliyaperumal, Deepa; Mantur, Neha Girish; and Rathikota; Nishanth Datta.
The application for the patent was published on August 14, 2026, under issue no. 33/2026.
Abstract: The present disclosure provides a system (100) and a method (300) for secure and robust diagnostic classification of dermoscopic skin-lesion images. An ingestion unit (106) loads dermoscopic images in successive batches and releases a loaded batch before a subsequent batch loads, and a feature unit (112) derives a hybrid feature vector (114) combining texture (112-2), shape (112-4) and color (112-6) descriptors with a patient clinical-metadata descriptor (112-8). Base classifiers (204) yield class predictions, a meta-classifier (206) yields a secure lesion diagnosis and a confidence score, a detection unit (208) raises a suspicion flag through a logical OR operation (208-8), and a mitigation unit (210) returns a retrained base classifier. Unlike conventional arrangements, the base classifiers (204) stay shared across secure diagnosis, cyber-attack detection, mitigation and recovery for dermoscopy screening in resource-limited clinical environment, thereby eliminating the need for separate cybersecurity models while improving computational efficiency and secure diagnostic performance.
Disclaimer: Curated by HT Syndication.
