Details

Autor(en) / Beteiligte

• Gola, Abhinav
• Panesar, Sonia
• Sharma, Aradhna
• Ananthakrishnan, Gayathri
• Singal, Gaurav
• Mukhopadhyay, Debajyoti
• Garg, Deepak
• Sarangapani, Jagannathan
• Wong, Kit
• Gupta, Suneet Kumar

Titel

MaskNet: Detecting Different Kinds of Face Mask for Indian Ethnicity

Ist Teil von

• Advanced Computing, p.492-503

Ort / Verlag

Singapore: Springer Singapore

Link zum Volltext

Quelle

Alma/SFX Local Collection

Beschreibungen/Notizen

• The COVID-19 pandemic has rendered social distancing and use of face masks as an absolute necessity today. Coming out of the epidemic, we're going to see this as the new normal and therefore most workplaces will require an identification system to permit employees based on the compliance of protocols. To ensure minimal contact and security, automatic entrance systems need to be employed in workplaces and institutions. For the implementation of such systems, we have investigated the performance of three object detection algorithms, namely SSD MobileNet V2, YOLO v3 and YOLO v4 in the context of real-time face mask detection. We conducted training and testing of these algorithms on our dataset focusing on various type of masks in the Indian community. We have exhibited in this paper that YOLOv4 transcends both YOLO v3 and SSD MobileNet V2 in sensitivity and precision and thus has a major use case in building AI identification systems.