Aug 2026· JIPI (Jurnal Ilmiah Penelitian dan Pembelajaran Informatika)· Vol 11, pp. 1810-1818· 0 citations
TL;DR
YOLOv12-N model, released in February 2025, which was designed using an efficient area attention-based architecture and a structured R-ELAN architecture, successfully combines com-putational efficiency with high accuracy and could be used in automated transportation systems.
Abstract
Transportation has become a basic necessity, enabling people to carry out their activities efficiently. One of the most widely used land transport sub-sectors is motorcycles. According to land transportation statistics, the number of motorcycles was 133.05 million from 2019 to 2023, with an annual growth rate of 4.22%. The most common viola-tion is not wearing a helmet. To prevent helmet violations, a system capable of automatically detecting vehicle images is needed. Computer systems play a key role in traffic monitoring. YOLO is an algorithm that uses a single convolutional neural network to detect objects in visual data, including pre-processing, training, and detection. This study evaluates the performance of the YOLOv12-N model, released in February 2025, which was designed using an efficient area attention-based architecture and a structured R-ELAN architecture, along with integrating FlashAttention. YOLOv12-N successfully combines com-putational efficiency with high accuracy. This study compares YOLOv12-N's performance with that of YOLOv8-N, which served as the baseline. The model was developed using Python in the Jupyter IDE. A dataset of 1,400 images was used in the study, showing that YOLOv12-N achieved an mAP@ IoU 0.5 of 68.13%, an mAP@ IoU 0.5:0.95 of 37.24%, a precision of 80.91%, and an F1 score of 73.42%. These results surpass those of YOLOv8-N: mAP@ IoU 0.5 of 65.26%, mAP@ IoU 0.5:0.95 of 35.59%, precision of 66.59%, and F1-score of 71.13%. Due to its stable performance and high precision, YOLOv12-N could be used in automated transportation systems. This discussion emphasizes the importance of developing AI-based systems.
Violations of helmet use regulations by motorcyclists in the Universitas Riau campus area still frequently occur. Supported by a preliminary survey indicating that 98.6% of 149 respondents perceive campus supervision as highly lenient, this lack of direct supervision creates a gap where students tend to ignore safety r...
Muhammad Ikram Syafwan, T. Y. Hadiwandra· JOURNAL OF APPLIED INFORMATI...· 0 citations
The number of individuals killed or hurt in road traffic and injuries can be significantly reduced if more persons
buckle up. Seatbelt detection is one of essential task in transportation system. However human monitoring of person -
seatbelt use from roadside or in vehicle camera error prone and labour intensive is req...
Amani Y. Noori, Huda H. Ali· International Journal of Adv...· 0 citations
Helmet compliance is an important requirement in smart transportation and industrial safety settings because a failure in detection in time and/or errors can cause severe injuries to the head and loss of life. The performance of one stage DL detectors in real-time object detection in the challenging visual scenarios ha...
B. S. Kumar, S. Veerla· Journal of Artificial Intell...· 0 citations
Electric two-wheelers are widely used in urban and campus traffic, but unsafe behaviors such as missing helmets and abnormal handlebar grasping remain difficult to detect continuously from road-scene images. This paper proposes a lightweight YOLOv8n-based detector for five-class dangerous-riding behavior detection of e...
Jia-Hao Chen, Xue Li, Bao-Jiang Ye et al.· 3rd International Conference...· 0 citations
Road safety has become a major concern due to the increasing number of traffic accidents caused by riders not wearing helmets. Traditional traffic monitoring systems mainly depend on manual supervision by traffic police, which is time-consuming, labour-intensive, and unable to provide continuous monitoring. As a result...
M. R. Chandra, R. V. Subash, P. S. Krishna et al.· International Conference on...· 0 citations
This paper addresses the limitations of traditional manual vehicle verification systems used in parking access control, which are often inefficient, time-consuming, and susceptible to human error. The study employs a three-stage deep learning based Automatic Number Plate Recognition (ANPR) framework using YOLOv8 for li...
Saroj Gautam, Janak Bikram Rana Magar, K. Chaudhary et al.· Journal of Hillside College...· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.