Real-Time Object Detection

Mrs. Annapurna Bhavani Koduri1, Annamreddi Nagesh2, Marada Padma Priya3,Sirasapalli Bhavani
Sankar4, Veugula Yaswanth5 1 Assistant Professor, Computer Science and Engineering, Visakha Institute
of Engineering & Technology(A), Narava, Visakhapatnam, India. 2,3,4 & 5 are the Students of Computer
Science and Engineering (Data Science) Vishaka Institute of Engineering and Technology(A), Completed
project under the guidance of Mrs. ANNAPURNA BHAVANI KODURI, Assistant Professor, Dept of CSE,
Vishaka Institute of Engineering and Technology(A).

ABSTRACT: The "Real-Time Object Detection " automates surveillance by replacing manual observation with an intelligent response framework using Python and YOLOv8. The system processes live video through a continuous pipeline, analyzing frames in a single regression task to identify objects and threats with low latency, thus eliminating human error. Built on a high-performance stack, the architecture integrates Ultralytics YOLOv8 for detection, OpenCV for video management, and Flask to serve a real-time dashboard. Crucially, the Pyttsx3 library provides offline text-to-speech alerts, enabling audible communication without an internet connection.The final platform delivers comprehensive security through visual and auditory intelligence. Upon activation, users view a live stream with dynamic bounding boxes and accuracy scores. A key feature is the dual-alert mechanism, pairing visual markers with instant voice notifications. The resulting dashboard offers a stable monitoring environment with persistent detection logs, proving that merging deep learning and web technologies creates a proactive, scalable security solution..
KEYWORD’S: Artificial Intelligence, YOLOv8, Object Detection, Computer Vision, Voice Alert, Flask, Real-TimeSurveillance.