MUMBAI, India, Aug. 12 -- Intellectual Property India has published a patent application (202611068313 A) filed by Manipal University Jaipur on June 01, 2026, for A Hybrid System For Extreme Compression And Ai-Assisted Reconstruction Of Images Over Ultra-Low Bandwidth Communication Channels.
Inventors include Utkrisht Thakur; Saihajveer Singh; and Jeyakrishnan Venugopal.
The application for the patent was published on August 07, 2026, under issue no. 32/2026.
Abstract: The present invention relates to a hybrid image compression and decompression system for ultra-low bandwidth image transmission capable of achieving compression ratios exceeding 1000:1 while preserving perceptual recognizability and structural fidelity. The system performs RGB standardization and LANCZOS-based downscaling prior to compression. Pixel data is compressed using Discrete Cosine Transform based coding for JPEG images and AV1 intra-frame coding for AVIF images along with quantization and entropy coding. Semantic metadata comprising edge maps, colour abstraction data, and luminance characteristics is generated to preserve structural and perceptual information. An adaptive compression control mechanism maintains predetermined file size limits. The compressed pixel data and metadata are packaged using ZIP and DEFLATE compression for efficient transmission. During decompression, metadata-guided reconstruction, structural alignment, colour restoration, Real-ESRGAN based super-resolution enhancement, and artifact correction improve visual quality over ultra-low bandwidth communication channels.
Disclaimer: Curated by HT Syndication.
