Umar Masud, Momin Siddiqui, Mohd. Sadiq, Sarfaraz Masood
This repository contains the official PyTorch implementation of SCS-Net: An efficient and practical approach towards Face Mask Detection. SCS-Net is a lightweight deep learning model with just 0.12M parameters which is a reduction of upto 496 times as compared to the state-of-the-art models designed for the task of facial mask detection.
We also provide a new dataset - Enhanced FMLD 2.0, which contains a large set of 25,296 synthetically designed incorrect face mask images. It is the first of its kind of data to be proposed with equal diversity and quantity. The data can be downloaded from Kaggle.
If you use this work for research or project purposes, then please credit the authors.
@article{MASUD20231878,
title = {SCS-Net: An efficient and practical approach towards Face Mask Detection},
journal = {Procedia Computer Science},
volume = {218},
pages = {1878-1887},
year = {2023},
note = {International Conference on Machine Learning and Data Engineering},
issn = {1877-0509},
doi = {https://doi.org/10.1016/j.procs.2023.01.165},
author = {Umar Masud and Momin Siddiqui and Mohd. Sadiq and Sarfaraz Masood}, }
We refer to code modules of SCS-Layer and SE-Net for our work. Thanks to their awesome work!