MUMBAI, India, Sept. 22 -- Intellectual Property India has published a patent application (202641110650 A) filed by Vallurupalli Nageswara Rao Vignana Jyothi Institute Of Engineering And Technology on September 15, 2026, for An Intelligent Artificial Intelligence-Based System And Method For Real-Time Video Background Removal, Object Detection, And Visual Scene Analysis Using Deep Learning.
Inventor includes Mrs. M. Swapna Kumari.
The application for the patent was published on September 18, 2026, under issue no. 38/2026.
Abstract: ABSTRACT [0020] A two-stage vision pipeline is introduced that first removes background clutter using DeepLabV3 semantic segmentation and then performs object detection with YOLOv8. The DeepLabV3 model generates precise foreground masks that isolate relevant objects from distracting surroundings, producing cleaner input for the subsequent detection stage. YOLOv8, fine-tuned on a custom collection of images containing persons, cars, dogs and cats, then locates and classifies the remaining objects with improved precision. Experiments conducted on both images and video sequences demonstrate that eliminating background noise reduces false detections and raises overall accuracy. Processed results, including cropped object instances and visual alerts, are stored and displayed through a simple interface. The modular design supports real-world applications in surveillance, autonomous systems and visual inspection where reliable separation of objects from complex backgrounds is required.
Disclaimer: Curated by HT Syndication.
