MUMBAI, India, Nov. 2 -- Intellectual Property India has published a patent application (202641125711 A) filed by Karpagam Academy Of Higher Education; Karpagam College Of Engineering; Vasanthakumar S; and Saravanan D on October 21, 2026, for Object Detection Using Deep Learning For Blind People.
Inventors include Vasanthakumar S; and Saravanan D.
The application for the patent was published on October 30, 2026, under issue no. 44/2026.
Abstract: Visual impairment significantly affects an individual’s ability to perceive and interact with the surrounding environment, often limiting independent mobility and increasing safety risks. Despite the availability of traditional assistive tools such as white canes and guide dogs, these solutions provide limited environmental information and lack object-level awareness. With rapid advancements in artificial intelligence and computer vision, there is a growing opportunity to develop intelligent assistive systems capable of interpreting visual data and communicating it in an accessible form. This project presents a real-time object detection and audio announcement system based on deep learning techniques to support visually impaired individuals in understanding their surroundings. The proposed system utilizes a camera to continuously capture live video streams of the environment. The captured frames are processed using a deep learning–based object detection algorithm, specifically the YOLO (You Only Look Once) model, which is known for its high detection speed and accuracy. YOLO enables the identification and classification of multiple objects within a single frame, making the system suitable for real-time applications. OpenCV is employed for video acquisition, frame preprocessing, and system integration, ensuring efficient handling of visual data under varying lighting and environmental conditions. Once objects are detected, their corresponding labels are extracted and transformed into meaningful textual information. This information is then converted into audible speech using text-to-speech technology such as Google Text-to-Speech. The generated audio output is delivered through speakers or headphones, allowing users to receive immediate verbal feedback about nearby objects without requiring any visual interaction. This auditory guidance enhances situational awareness and assists users in making informed navigation and safety decisions. The system is designed to operate continuously and autonomously, providing real-time object recognition and announcements as the user moves through different environments. It is adaptable to both indoor and outdoor scenarios and can be deployed on various platforms, including personal computers, embedded systems, and wearable devices. The use of open-source frameworks and pre-trained models makes the solution cost-effective and accessible, enabling widespread adoption in real-world applications. In addition to assisting visually impaired individuals, the proposed system has potential applications in areas such as smart surveillance, robotics, human–computer interaction, and safety monitoring, where real-time object detection and audio feedback are essential. The modular architecture of the system allows for future enhancements, including distance estimation, obstacle prioritization, multilingual speech output, and integration with navigation systems. Overall, this project demonstrates an effective combination of deep learning, computer vision, and speech synthesis technologies to create an intelligent assistive solution. By providing real-time object detection with verbal announcements, the system aims to improve environmental awareness, enhance personal safety, and promote greater independence for visually impaired users, contributing meaningfully to the field of assistive technology.
Disclaimer: Curated by HT Syndication.
