This paper presents a YOLOv8-based deep learning system for automatic Fire and Smoke detection and localization in images and video frames. A custom dataset is prepared with bounding-box annotations for Fire and Smoke classes. The proposed workflow includes data collection, annotation, preprocessing, model training, validation, and performance evaluation. The requested project draft reports an illustrative mAP@0.5 of 85%, with Precision of 85%, Recall of 83%, and F1-score of 84%. These values are draft target values and should be replaced by actual model results before formal publication. The proposed system can support CCTV-based monitoring in educational campuses, laboratories, warehouses, industries, public buildings, and other camera-monitored environments.

Smiritika SV, Pooja A, Sruthin G S, Sivamani A, Vidyabharathi T,"YOLOv8-Based Fire and Smoke Detection Using Deep Learning" International Journal of Advanced Multidisciplinary Research and Educational Development, V2(4): Page(413-417) July - August 2026. ISSN: 3107-6513. www.ijamred.com. Published by Scientific and Academic Research Publishing.