AI-Driven Object and Scene Recognition System for Blind Assistance

Abstract: The rapid advancement of Internet of Things (IoT) and Artificial Intelligence (AI) technologies has enabled the development of intelligent systems capable of perceiving and interpreting real-world environments. This project presents an Intelligent Visual Assistant that integrates an ESP32-CAM module with a server based AI object detection framework to provide real-time visual understanding and voice-based feedback. The ESP32-CAM captures images and transmits them wirelessly to a Flask-based server, where the images are processed using the YOLOv8 object detection algorithm to identify objects present in the scene. The detected information is converted into audible speech using a Text-to-Speech (TTS) engine and delivered to the user through a speaker or Bluetooth-enabled device. The proposed system adopts a hybrid edge–server architecture that balances low-cost hardware with high-performance AI processing. Experimental results demonstrate reliable object detection accuracy with low response latency under local network conditions, making the system suitable for real-time applications. The solution is particularly beneficial for assistive technologies for visually impaired individuals, smart surveillance etc. Key Words: ESP32-CAM, Object Detection, YOLOv8, Internet of Things (IoT), Text-to-Speech (TTS), Real-Time Image Processing, Assistive Technology, Embedded Systems, Wireless Communication