MUMBAI, India, Oct. 26 -- Intellectual Property India has published a patent application (202611111768 A) filed by Ims Engineering College, Ghaziabad on September 17, 2026, for A System For Real-Time Video Summarization Using Adaptive Deep Feature Prioritization And Event- Consistent Summary Generation.
Inventors include Prof. Dr. Anil K. Ahlawat; Dr. Vikash Goel; and Dr. Ramesh Kumar Verma.
The application for the patent was published on October 23, 2026, under issue no. 43/2026.
Abstract: [048] The present invention discloses a system for real-time video summarization using adaptive deep feature prioritization and event-consistent summary generation. The system comprises a video acquisition interface, a temporal frame buffer, a deep feature extraction accelerator, an adaptive saliency scoring engine, an event continuity analyser, a redundancy suppression engine, a summary generation module, and an output interface. The system receives an incoming video stream, converts the stream into normalized frame units, extracts compact visual feature embeddings, assigns adaptive importance scores, links related frames into event chains, removes redundant content, and generates a concise summary output in real time. The generated summary may include a shortened video, key-frame sequence, storyboard, event-indexed clip list, or metadata-enriched summary. The invention reduces video review time, storage requirement, and transmission load while preserving meaningful visual events, temporal continuity, and contextual relevance of the original video. Accompanied Drawing [FIGS. 1-2]
Disclaimer: Curated by HT Syndication.
