MUMBAI, India, July 24 -- Intellectual Property India has published a patent application (202611057138 A) filed by Indian Institute Of Technology Banaras Hindu University, Varanasi on May 05, 2026, for An Intelligent Hyperspectral Image Classification System And Method.
Inventors include Rajeev Srivastava; and Rajat Kumar Arya.
The application for the patent was published on July 17, 2026, under issue no. 29/2026.
Abstract: The intelligent hyperspectral image classification system (100) comprising a processor (101) including input module (1) to receive hyperspectral image data comprising spectral bands, spatial patch extraction module (2) configured to extract a spatial-spectral patch corresponding to a target pixel, spectral token encoding module (3) configured to divide spectral bands into spectral tokens and encode the spectral tokens to reduce redundancy, token interaction statistics module (4) configured to compute statistical representations of the spectral tokens, kernel interaction generator (5) configured to dynamically generate convolution kernels based on the computed statistical representations, a spatial-spectral mixing module (6) to apply the dynamically generated convolution kernels to the spatial- spectral patch to generate fused feature representations, event-based attention refinement module (7) to activate attention based on a divergence condition associated with the fused feature representations and classification module (8) comprising classifiers to generate prediction outputs and perform confidence-weighted aggregation to produce a final classification output. Figure 2 on sheet no. 2 of the drawings may accompany the abstract when published.
Disclaimer: Curated by HT Syndication.
