Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{200279,
        author = {Devaki Lahari and B. Raveendra Naick and B. Rama sivaji and Ch. Manikantachari and D.V. Siva Rama Krishna},
        title = {Accident Detection using YOLO + CNN Hybrid Model},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {12},
        number = {12},
        pages = {703-711},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=200279},
        abstract = {Road accidents are a major global concern, requiring timely detection to improve emergency response and reduce fatalities. Traditional monitoring methods based on manual observation are often slow and inefficient [1]. This paper proposes a hybrid deep learning approach for automated accident detection by combining YOLOv8 for object detection and a Convolutional Neural Network (CNN) for classification.
In the proposed system, YOLOv8 first detects vehicles in an image and extracts relevant regions, which are then classified by a CNN (MobileNetV2) as either “Accident” or “No Accident.” The model is trained and evaluated on a CCTV Accident Detection dataset consisting of approximately 989 images. Compared to conventional CNN-based approaches that process entire images, the proposed method focuses on relevant regions, improving real- world accuracy to around 80% [2].
The system also provides bounding boxes for better localization and interpretability. Overall, the hybrid model enhances detection accuracy and efficiency, making it suitable for intelligent traffic monitoring systems, with future potential for real-time applications and automated alert systems.},
        keywords = {Accident Detection, YOLOv8, CNN, Deep Learning, Computer Vision, Hybrid Model},
        month = {May},
        }