The IRMA Community
Newsletters
Research IRM
Click a keyword to search titles using our InfoSci-OnDemand powered search:
|
On the Detection of Faces With Masks Using Tiny YOLOv7 Algorithm
Abstract
Today every individual is expected to wear a mask, which poses a new challenge to security and surveillance of individuals for any governing body. Though notable work has been done in the area of face mask detection, there still exists a bottleneck of fast detection. Additionally, the complexity of features, size of frames, and inhomogeneity of data poses a challenge to achieve a model with high accuracy. And law offenders are quick to exploit this opportunity to their advantage. Through this work, the aim is to propose a system that combines Tiny YOLOv7 and Jestson Nano which is able to detect faces with or without mask based on the recently introduced Tiny YOLOv7 algorithm. The proposed system was able to achieve a mAP of 55.94% and an average IoU of 53.70%. The average precision achieved for people with masks was 83.80% and 79.67% for specific detection of the mask region. The model uses a total of 5.527 BFLOPs and was able to achieve an average FPS of 71.8, which ensures a higher throughput leading to a faster model both in terms of training and detection.
Related Content
|
G. Boopathy, Balaji Ganesan, P. Sivaprakasam, T. Kumaran.
© 2026.
42 pages.
|
|
G. Prasad.
© 2026.
14 pages.
|
|
Kishorebabu Dasari, Sujana Parry, Srinivas Mekala.
© 2026.
30 pages.
|
|
Chikesh Ranjan, Jonnalagadda Srinivas, P. S. Balaji, Kaushik Kumar.
© 2026.
24 pages.
|
|
G. Ananthi, S. Mehala Shevani, P. Priyadharshini Devi.
© 2026.
24 pages.
|
|
G. Prasad, Snehal Malik, Aadya Gupta, Yash Nigam.
© 2026.
26 pages.
|
|
Dhirendra Patel, M. L. Azad.
© 2026.
36 pages.
|
|
|