Real-time weapon detection using deep learning

In recent years, increasing security threats, particularly in crowded areas and strategically important facilities, have made the rapid and reliable detection of weapons essential. Traditional security measures are often insufficient, and surveillance conducted by the human eye is both time-consuming and prone to errors. Deep learning–based object detection models enable the real-time classification and tracking of different types of weapons. This study explores the effectiveness of various real-time object detectors for the problem of classifying different weapon types. To that end, the performance of YOLOv9, YOLOv10, YOLOv11, YOLOX, and YOLOR models were evaluated on a balanced dataset of 15,387 images across 5 weapon classes. All models are trained with 5-fold cross-validation and YOLOR achieved the best overall performance. Results show YOLOR's suitability for high-precision, critical applications, while YOLOX offers advantages for recall-focused, resource-constrained deployments. The experiments highlighted the potential of real-time AI-based object detectors for enhancing security systems.