MUMBAI, India, Feb. 6 -- Intellectual Property India has published a patent application (202641007105 A) filed by Sivakrishna Reddy Vundela; and Priyanka Ankireddy, Kadapa, Andhra Pradesh, on Jan. 24, for 'a method and apparatus for early detection of road accidents using deep learning on video streams.'
Inventor(s) include Sivakrishna Reddy Vundela; and Priyanka Ankireddy.
The application for the patent was published on Feb. 6, under issue no. 06/2026.
According to the abstract released by the Intellectual Property India: "The present invention relates to a method and apparatus for early detection of road accidents using deep learning on video streams. The invention provides an intelligent video analytics system capable of predicting and detecting imminent or potential road accidents by analyzing real-time or recorded video feeds obtained from traffic surveillance cameras, vehicle-mounted cameras, or other video sensing devices. The system preprocesses the captured video streams to normalize frames and enhance visual quality, followed by extraction of spatial and temporal features representing vehicle motion, trajectories, interactions, and abnormal driving behaviors. A deep learning-based prediction engine, comprising convolutional neural networks in combination with temporal learning models, processes the extracted features to compute an accident risk score indicative of a high-probability accident scenario. The risk score is evaluated against predefined or adaptive thresholds to determine the likelihood of an impending accident. Upon exceeding the threshold, the system generates early warning alerts that may be transmitted to traffic management centers, emergency response units, nearby vehicles, or authorized monitoring systems through wired or wireless communication networks.The invention enables proactive accident detection prior to or at the early stage of occurrence, thereby reducing response time, improving road safety, minimizing casualties, and enhancing traffic management efficiency. The proposed method and apparatus are scalable, adaptable to varying traffic environments, and suitable for deployment in urban roads, highways, and smart transportation infrastructures."
Disclaimer: Curated by HT Syndication.
