Detecting pistol-type weapons using convolutional networks with YOLO-like architecture and stereoscopy
Abstract
Security cameras and video surveillance systems play a crucial role in ensuring public safety. However, the increasing accessibility of pistol-type firearms contributes to growing concerns about insecurity. Early detection of these weapons is of utmost importance to prevent potential accidents. This study aims to develop a real-time stereoscopic vision system capable of accurately detecting pistol-type objects and determining their distance with high confidence. The approach combines a convolutional neural network (CNN) architecture with a YOLO-type algorithm, utilizing transfer learning, and incorporates an algorithm for stereoscopic distance estimation. The presented system achieves an accuracy of 92.2 % with an Intersection over Union (IoU) value of 0.6. Moreover, the average distance estimation error within a range of 3 meters is only 9.3 centimeters.
Downloads
References
Arballo, J. J. P., Diaz, S. M., Liera, M. A. C., y Taylor, J. E. L. (2022). Detec- cion de cambio en superficie costera mediante la segmentacion de imagenes aereas utilizando redes neuronales convolucionales. Padi Boletin Cientifico de Ciencias Basicas e Ingenierias del ICBI, 10:136–144.
Chavelas, S. M. y Diaz, S. M. (2019). Algoritmo paralelo en gpu para el rastreo de armas de fuego tipo pistola tesis.
Deepthi, T., Gaayathri, R., y Shanthosh, S. (2018). Firearm recognition using convolutional neural network. academia.edu.
Flitton, G., Breckon, T. P., y Megherbi, N. (2013). A comparison of 3d inter- est point descriptors with application to airport baggage object detection in complex ct imagery. Pattern Recognition, 46:2420–2436.
Gesick, R., Saritac, C., y Hung, C. C. (2009). Automatic image analysis pro- cess for the detection of concealed weapons. ACM International Conference Proceeding Series.
Grega, M., Matiolanski, A., Guzik, P., y Leszczuk, M. (2016). Automated de- tection of firearms and knives in a cctv image.
INEGI (2021). Datos preliminares revelan que en 2020 se registraron 36579 homicidios.
Jocher, G. y Ultralytics (2023). Github yolov8 - ultralytics. Accessed on February 13, 2023
Liu, F., Zhao, H., y Liu, W. (2022). Improved garment detection algorithm based on yolov5. pp. 1054–1058.
Olmos, R., Tabik, S., y Herrera, F. (2018). Automatic handgun detection alarm in videos using deep learning. Neurocomputing, 275:66–72.
RangeKing (2023). Brief summary of yolov8 model structure. Accessed on February 13, 2023.
Redmon, J., Divvala, S., Girshick, R., y Farhadi, A. (2016). You only look once: Unified, real-time object detection. pp. 779–788.
Remondino, F. ., Fraser, C., y Remondino, F. (2006). Digital camera calibration methods. International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, XXXVI:266–272.
Strbac, B., Gostovic, M., Lukac, Z., y Samardzija, D. (2020). Yolo multi- camera object detection and distance estimation. 2020 Zooming Innovation in Consumer Technologies Conference, ZINC 2020, pp. 26–30.
SUSMITHA, D. y KUMAR, S. V. (2023). Weapon detection using artificial intelligence and deep learning for security applications. Journal of Engineering Sciences, 14(01).
Veit, A., Matera, T., Neumann, L., Matas, J., y Belongie, S. (2016). Coco-text: Dataset and benchmark for text detection and recognition in natural images. arXiv preprint arXiv:1601.07140
Copyright (c) 2023 Aarón Schcolnik-Elias, Saúl Martínez-Díaz, Jorge Enrique Luna-Taylor, Iliana Castro-Liera
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
