SqueezeMaskNet: Real-Time Mask-Wearing Recognition for Edge Devices
This paper presents SqueezeMaskNet, a lightweight convolutional neural network designed for real-time recognition of proper and improper mask usage. The model classifies four categories: masks worn correctly, masks covering only the mouth, masks not covering, and no mask. SqueezeMaskNet integrates s...
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2025-01-01
|
| Series: | Big Data and Cognitive Computing |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2504-2289/9/1/10 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832589127277608960 |
|---|---|
| author | Gibran Benitez-Garcia Lidia Prudente-Tixteco Jesus Olivares-Mercado Hiroki Takahashi |
| author_facet | Gibran Benitez-Garcia Lidia Prudente-Tixteco Jesus Olivares-Mercado Hiroki Takahashi |
| author_sort | Gibran Benitez-Garcia |
| collection | DOAJ |
| description | This paper presents SqueezeMaskNet, a lightweight convolutional neural network designed for real-time recognition of proper and improper mask usage. The model classifies four categories: masks worn correctly, masks covering only the mouth, masks not covering, and no mask. SqueezeMaskNet integrates seamlessly with existing face detection systems, removing the need for retraining. We propose using Fire modules for efficiency, along with attention mechanisms like efficient channel attention (ECA) and squeeze-and-excitation (SE) blocks for improved feature refinement. SqueezeMaskNet achieved 96.7% accuracy on the challenging FineFM test set and ran at 297 FPS on a GPU and up to 96 FPS on edge devices like a Jetson Orin NX. We also introduced ImproperTFM, a subset of real-world images focusing on improper mask usage, which enhanced the model accuracy when combined with FineFM data. Comparative experiments demonstrated SqueezeMaskNet’s superior performance, efficiency, and adaptability compared to MobileNet and EfficientNet, making it a practical solution for mask-wearing recognition across various devices and settings. |
| format | Article |
| id | doaj-art-8066fb3dfdd94fc79246bf034134c490 |
| institution | Kabale University |
| issn | 2504-2289 |
| language | English |
| publishDate | 2025-01-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Big Data and Cognitive Computing |
| spelling | doaj-art-8066fb3dfdd94fc79246bf034134c4902025-01-24T13:22:32ZengMDPI AGBig Data and Cognitive Computing2504-22892025-01-01911010.3390/bdcc9010010SqueezeMaskNet: Real-Time Mask-Wearing Recognition for Edge DevicesGibran Benitez-Garcia0Lidia Prudente-Tixteco1Jesus Olivares-Mercado2Hiroki Takahashi3Graduate School of Informatics and Engineering, The University of Electro-Communications, Chofugaoka 1-5-1, Chofu-shi 182-8585, JapanInstituto Politecnico Nacional, ESIME Culhuacan, Mexico City 04440, MexicoInstituto Politecnico Nacional, ESIME Culhuacan, Mexico City 04440, MexicoGraduate School of Informatics and Engineering, The University of Electro-Communications, Chofugaoka 1-5-1, Chofu-shi 182-8585, JapanThis paper presents SqueezeMaskNet, a lightweight convolutional neural network designed for real-time recognition of proper and improper mask usage. The model classifies four categories: masks worn correctly, masks covering only the mouth, masks not covering, and no mask. SqueezeMaskNet integrates seamlessly with existing face detection systems, removing the need for retraining. We propose using Fire modules for efficiency, along with attention mechanisms like efficient channel attention (ECA) and squeeze-and-excitation (SE) blocks for improved feature refinement. SqueezeMaskNet achieved 96.7% accuracy on the challenging FineFM test set and ran at 297 FPS on a GPU and up to 96 FPS on edge devices like a Jetson Orin NX. We also introduced ImproperTFM, a subset of real-world images focusing on improper mask usage, which enhanced the model accuracy when combined with FineFM data. Comparative experiments demonstrated SqueezeMaskNet’s superior performance, efficiency, and adaptability compared to MobileNet and EfficientNet, making it a practical solution for mask-wearing recognition across various devices and settings.https://www.mdpi.com/2504-2289/9/1/10proper mask-wearing recognitionlightweight neural networksSqueezeMaskNetreal-time inferenceFineFM dataset |
| spellingShingle | Gibran Benitez-Garcia Lidia Prudente-Tixteco Jesus Olivares-Mercado Hiroki Takahashi SqueezeMaskNet: Real-Time Mask-Wearing Recognition for Edge Devices Big Data and Cognitive Computing proper mask-wearing recognition lightweight neural networks SqueezeMaskNet real-time inference FineFM dataset |
| title | SqueezeMaskNet: Real-Time Mask-Wearing Recognition for Edge Devices |
| title_full | SqueezeMaskNet: Real-Time Mask-Wearing Recognition for Edge Devices |
| title_fullStr | SqueezeMaskNet: Real-Time Mask-Wearing Recognition for Edge Devices |
| title_full_unstemmed | SqueezeMaskNet: Real-Time Mask-Wearing Recognition for Edge Devices |
| title_short | SqueezeMaskNet: Real-Time Mask-Wearing Recognition for Edge Devices |
| title_sort | squeezemasknet real time mask wearing recognition for edge devices |
| topic | proper mask-wearing recognition lightweight neural networks SqueezeMaskNet real-time inference FineFM dataset |
| url | https://www.mdpi.com/2504-2289/9/1/10 |
| work_keys_str_mv | AT gibranbenitezgarcia squeezemasknetrealtimemaskwearingrecognitionforedgedevices AT lidiaprudentetixteco squeezemasknetrealtimemaskwearingrecognitionforedgedevices AT jesusolivaresmercado squeezemasknetrealtimemaskwearingrecognitionforedgedevices AT hirokitakahashi squeezemasknetrealtimemaskwearingrecognitionforedgedevices |