VISTA: Vision-Integrated Smart Tracking Assistant

Suhas Sreehari, Arya R Menon, D. Malavika, C K Navaneeth, R M Devrag, Manazhy Rashmi · 2025

Indoor navigation can be a serious challenge for visually impaired persons, so they require rather sophisticated assistive technology. VISTA: Vision-Integrated Smart Tracking Assistant is an up-and-coming technology that will address indoor way finding, object identification, real-time hazard detection, and scheduling for blind users. In addition, through computer vision, machine learning, and TTS technologies, this system could give an experience of being safer and a lot more independent. The primary features of VISTA include, but are not limited to, realtime object detection, hazard detection, voice-guided navigation, and dynamic reading of text. The system utilizes a camera-based system to find any obstacles and important landmarks while providing the user with auditory feedback in real time. In addition, a real-time reading module of signboard and danger signal has been integrated into the system to identify and read written text into speech thus warning the user for relevant announcements, room numbers, or warnings. An integrated keyboard-based schedule and reminder system has been produced for the benefit of daily users’ life management. This system, developed in Python, employs pyttsx 3 for offline text-to-speech conversion, the keyboard library for detection of user input, and the date time module for time-related retrieval. Users are also easily facilitated to their schedules, reminders, and notifications through very simple key press operations-the easy access and efficiency conferred on them. The system was tested to have real-world repeatability in accuracy, usability, and response time. Future enhanced features will include AI-powered obstacle detection, customized voice alerts, and cloud data synchronization for enhanced usability. System integrates QR based indoor navigation for precise and exact location tracking and management within buildings.

Read the paper · More papers on PaperTik