MUMBAI, India, June 30 -- Intellectual Property India has published a patent application (202441101602 A) filed by Manipal Academy Of Higher Education on December 21, 2024, for A System And Method For Real-Time Reflection And Distortion Removal On Edge Devices.
Inventor includes Christie Mathews.
The application for the patent was published on June 26, 2026, under issue no. 26/2026.
Abstract: The invention provides a system and method for real-time reflection and distortion removal using an optimized Autoencoder-based model for edge devices. It addresses challenges in image clarity, computational inefficiency, and latency by enabling local processing on resource-constrained devices such as smartphones and IoT systems. The method involves preprocessing input images, encoding and decoding them through the Autoencoder to eliminate distortions and reflections, and leveraging hardware acceleration for rapid inference. An API interface facilitates remote integration, while optional postprocessing enhances output quality. The system ensures efficient, low-latency image correction suitable for diverse applications, including augmented reality, autonomous systems, healthcare imaging, and smart city technologies. By combining advanced neural networks with edge computing, the invention enables real-time usability, privacy, and broad applicability in dynamic and resource-limited environments.
Disclaimer: Curated by HT Syndication.
