MUMBAI, India, July 7 -- Intellectual Property India has published a patent application (202641076801 A) filed by Vellore Institute Of Technology on June 21, 2026, for System And Method For Condition-Aware Severity Estimation In Dermatological Image Analysis Using A Feedback-Gated Probabilistic Dependency Network.
Inventors include Sivaprakash S; Reena Sri G; and Saktheshwaran T.
The application for the patent was published on July 03, 2026, under issue no. 27/2026.
Abstract: The present invention relates to a system and method for condition-aware severity estimation in dermatological image analysis, particularly for lip-related disorders. The system receives input lip images (101), performs image pre-processing (102), and extracts visual features through a feature extraction module (103) comprising a convolutional neural network backbone. A condition and severity classification module (104) including a lip condition detection module (105) generates condition probability values for multiple lip conditions. A Probabilistic Dependency Feedback Gate utilizes the condition probabilities to dynamically modulate extracted feature representations before severity estimation, thereby establishing an explicit dependency between condition detection and severity prediction. A Context-Aware Intensity Regression Module computes continuous severity intensity scores based on condition confidence, relative ranking, and cross-condition interaction terms. The intensity scores are mapped to interpretable severity levels and used to generate condition-specific recommendations. The invention improves severity estimation accuracy by incorporating condition interdependencies and provides an end-to-end framework suitable for dermatological assessment and telemedicine applications.
Disclaimer: Curated by HT Syndication.
