MUMBAI, India, June 30 -- Intellectual Property India has published a patent application (202641076431 A) filed by Koneru Lakshmaiah Education Foundation Deemed To Be University on June 19, 2026, for Hybrid Deep Learning Architecture For Context-Aware Visual Question Answering Using Visual Attention And Language Intelligence.
Inventors include Y Harika Devi; and Dr. N Chaitanya Kumar.
The application for the patent was published on June 26, 2026, under issue no. 26/2026.
Abstract: The present invention discloses a hybrid deep learning architecture for context-aware visual question answering by integrating visual attention mechanisms, language intelligence, multimodal feature fusion, and contextual reasoning techniques. The architecture receives visual input data and associated natural language queries, extracts visual representations, and identifies relevant visual regions using an attention-driven mechanism. A language intelligence module generates contextual linguistic embeddings to understand query semantics, while a multimodal fusion module combines visual and textual information to create a unified representation. A contextual reasoning module analyzes relationships between visual elements and linguistic concepts for generating accurate responses through an answer generation module. The invention improves visual understanding, semantic alignment, reasoning capability, and response accuracy, enabling applications in intelligent assistants, autonomous systems, healthcare analysis, surveillance, education platforms, and advanced human-computer interaction environments.
Disclaimer: Curated by HT Syndication.
