MUMBAI, India, July 24 -- Intellectual Property India has published a patent application (202641063212 A) filed by Manipal Academy Of Higher Education on May 19, 2026, for A System And Method For Asymmetric Dual Resolution Image Processing With Adaptive Feature Fusion.
Inventors include Pavithra K C; Preetham Kumar; Geetha M; and Sulatha V Bhandary.
The application for the patent was published on July 17, 2026, under issue no. 29/2026.
Abstract: The present invention relates to a system and method for dual resolution image processing using adaptive feature fusion. The method includes acquiring image data and extracting a region of interest corresponding to a physically calibrated predefined dimension, followed by generating asymmetric inputs comprising a full image at a reduced resolution and a region-specific image at a higher resolution. The full image and the region of interest are processed through separate pathways to obtain global and localized feature representations. The feature representations are combined using an adaptive weighting mechanism that dynamically assigns relative importance to each pathway. The combined features are further transformed to generate a unified representation for producing a classification output. The proposed approach enables efficient multi-scale feature extraction while achieving reduced per-inference computational complexity relative to conventional single-pathway uniform-resolution systems, without compromising diagnostic discriminative performance. FIG.2
Disclaimer: Curated by HT Syndication.
