A Comprehensive Analysis of Mask Detection using Convolutional Neural Networks (CNN) and Single Shot Multibox Detector (SSD) Approach

Kusmantoro, Adhi (2023) A Comprehensive Analysis of Mask Detection using Convolutional Neural Networks (CNN) and Single Shot Multibox Detector (SSD) Approach. In: 2023 International Conference on Sustainable Emerging Innovations in Engineering and Technology (ICSEIET).

Text A_Comprehensive_Analysis_of_Mask_Detection_Using_Convolutional_Neural_Networks_CNN_and_Single_Shot_M (1).pdf Download (1MB)

Abstract

The Covid-19 pandemic has emphasized the critical role of mask-wearing in reducing virus transmission. However, declining public awareness necessitates effective monitoring solutions. This study explores the Single Shot Multibox Detector (SSD) approach, a Convolutional Neural Networks (CNN)-based object detection method, for mask detection. Specifically, we investigate the performance of the SSD model with the ResNet50 architecture as the base network. Through the analysis of a dataset comprising 1000 images of individuals wearing and not wearing masks, the SSD model with ResNet50 achieves an impressive accuracy of 98.23%, surpassing the VGG16-based model's accuracy of 93.56%. These findings underscore the superiority of the CNN-based SSD model with ResNet50 for mask detection. Moreover, they have significant implications for real-time monitoring and enforcing mask-wearing protocols, emphasizing the importance of public awareness and adherence. Leveraging advanced AI techniques, such as CNN-based object detection with SSD, strengthens public health initiatives. The development of robust and efficient mask detection systems contributes to creating a safer environment amid the ongoing Covid-19 pandemic. This research highlights the value of harnessing these advanced technologies to effectively address public health crises. By employing CNN-based SSD with ResNet50, society can enhance its ability to mitigate risks and ensure public safe

Item Type:	Conference or Workshop Item (Paper)
Subjects:	T Technology > TK Electrical engineering. Electronics Nuclear engineering
Depositing User:	dosen upgris semarang
Date Deposited:	21 May 2024 07:19
Last Modified:	21 May 2024 07:19
URI:	http://eprints.upgris.ac.id/id/eprint/3273

Actions (login required)

View Item