Author(s)

XinLei Yang, JueXiao Chen^*

Affiliation(s)

School of Automotive Engineering, Tongji University, Shanghai 200092, China.

Corresponding Author

JueXiao Chen

ABSTRACT

To improve the efficiency of ambulance emergency task scheduling and address the issues of low efficiency and poor real-time performance in traditional manual entry of stretcher status, this paper proposes an automatic judgment method for ambulance emergency status based on deep learning. A self-built ambulance stretcher dataset is constructed by combining 567 images captured independently from ambulances and 545 images obtained via web crawling. Data augmentation techniques are used to enhance model robustness, and mainstream object detection algorithms such as TOOD, Faster R-CNN, and YOLOv8 are compared and analyzed. Experimental results show that YOLOv8 achieves an average precision (AP@0.5) of 0.779 at an IoU threshold of 0.5, with only 11.2 million parameters, significantly outperforming other models. This method can accurately and real-time determine the stretcher status, providing a reliable basis for emergency centers to dynamically dispatch ambulance resources. It effectively shortens the emergency response time and improves the success rate of emergency rescue.

KEYWORDS

Ambulance emergency status; Object detection; Deep learning; Data augmentation

CITE THIS PAPER

XinLei Yang, JueXiao Chen. Exploration of the detection method of the condition of patients on the ambulance stretcher. World Journal of Engineering Research. 2025, 3(1): 47-51. DOI: https://doi.org/10.61784/wjer3017.

REFERENCES

[1] Hu Jingchun, Liu Xiaomei, Yang Cheng, et al. Influence of Differences in Pre-hospital Emergency Cardiopulmonary Resuscitation Conditions on the Efficacy of Patients with Cardiac Arrest. Journal of Anhui Health Vocational & Technical College, 2013, 12(01): 27-28.

[2] Zhang Wenwu, Dou Qingli, Liang Jinfeng, et al. Government-led Public First Aid Training: The Practice in Bao'an, Shenzhen. Chinese Journal of Emergency Medicine, 2019(01): 126-128.

[3] S M, T R, N V, et al. Adaptive ambulance monitoring system using IOT. Measurement: Sensors, 2022, 24.

[4] Gao Wenxuan, Yang Xinjie. A Computation Offloading Scheme for Energy Consumption Optimization in Internet of Vehicles. Telecommunications Science, 2023, 39(10): 29-40.

[5] Papakipos Z, Bitton J. Augly: Data augmentations for robustness. arxiv preprint arxiv:2201.06494, 2022.

[6] Feng C, Zhong Y, Gao Y, et al. Tood: Task-aligned one-stage object detection. 2021 IEEE/CVF International Conference on Computer Vision (ICCV). IEEE Computer Society, 2021: 3490-3499.

[7] Ren S. Faster r-cnn: Towards real-time object detection with region proposal networks. arxiv preprint arxiv:1506.01497, 2015.

[8] Redmon J. You only look once: Unified, real-time object detection. Proceedings of the IEEE conference on computer vision and pattern recognition. 2016.

[9] Cai Z, Vasconcelos N. Cascade r-cnn: Delving into high quality object detection. Proceedings of the IEEE conference on computer vision and pattern recognition. 2018: 6154-6162.

[10] He K, Gkioxari G, Dollár P, et al. Mask r-cnn. Proceedings of the IEEE international conference on computer vision. 2017: 2961-2969.

[11] Redmon J. Yolov3: An incremental improvement. arxiv preprint arxiv:1804.02767, 2018.

[12] Terven J, Córdova-Esparza D M, Romero-González J A. A comprehensive review of yolo architectures in computer vision: From yolov1 to yolov8 and yolo-nas. Machine Learning and Knowledge Extraction, 2023, 5(4): 1680-1716.

[13] Li X, Wang W, Wu L, et al. Generalized focal loss: Learning qualified and distributed bounding boxes for dense object detection. Advances in Neural Information Processing Systems, 2020, 33: 21002-21012.

[14] Zhang S, Chi C, Yao Y, et al. Bridging the gap between anchor-based and anchor-free detection via adaptive training sample selection. Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. 2020: 9759-9768.