Manuscript Number : SHISRRJ24726
A Swin Transformer-Based Approach for Motorcycle Helmet Detection
Authors(2) :-Dr. K. Shanmugam, Kakimanu. Chandana By supporting enforcement actions, automated video surveillance-based helmet wear identification among motorbike riders possesses the capacity to increase traffic safety. In spite of this, there are numerous drawbacks to the existing detection methods. For example, they can't distinguish between several passengers or work well in complicated environments. In this study, we combine computer vision and machine learning to tackle the difficult challenge of automated helmet use monitoring. We suggest a technique called transformers that is grounded in models of deep neural networks. The Swin transformer's base version serves as the foundation for feature extraction, and for final detection, we integrate the Cascade Area-based Convolution Framework for Neural Networks (RCNN) with a Neck of the Feature Pyramid Network (FPN). Our proposed strategy's effectiveness is validated by extensive testing and compared with current methods. Our model's mean average precision (mAP) was 30.4. approach performs better than existing methods for detection.
Dr. K. Shanmugam Deep learning, intelligent transportation systems, transformers, motorbike safety, and helmet detection. Publication Details Published in : Volume 7 | Issue 2 | March-April 2024 Article Preview
Assistant Professor, Department of MCA, Annamacharya Institute of Technology & Sciences, Tirupati, Andhra Pradesh, India
Kakimanu. Chandana
Post Graduate, Department of MCA, Annamacharya Institute of Technology & Sciences, Tirupati, Andhra Pradesh, India
Date of Publication : 2024-04-30
License: This work is licensed under a Creative Commons Attribution 4.0 International License.
Page(s) : 35-42
Manuscript Number : SHISRRJ24726
Publisher : Shauryam Research Institute
URL : https://shisrrj.com/SHISRRJ24726