يك رويكرد يادگيري انتقالي با شبكه عصبي كانولوشنال براي تشخيص افراد داراي ماسك از روي تصاوير

A transfer learning approach with convolutional neural network for Face Mask Detection

ﺑﺎ ﺗﻮﺟﻪ ﺑﻪ ﻫﻤﻪﮔﻴﺮﻱ ﻭﻳﺮﻭﺱ ﮐﺮﻭﻧﺎ )ﮐﻮﻭﻳﺪ- 19( ﻭ ﺍﻧﺘﻘﺎﻝ ﺳﺮﻳﻊ ﺁﻥ ﺩﺭ ﺳﺮﺗﺎﺳﺮ ﺩﻧﻴﺎ، ﺟﻬﺎﻥ ﺑﺎ ﻳﮏ ﺑﺤﺮﺍﻥ ﺑﺰﺭﮒ ﺭﻭﺑﺮﻭ ﺷﺪﻩ ﺍﺳﺖ. ﺑﺮﺍﻱ ﺟﻠﻮﮔﻴﺮﻱ ﺍﺯ ﺷﻴﻮﻉ ﻭﻳﺮﻭﺱ ﮐﺮﻭﻧﺎ ﺳﺎﺯﻣﺎﻥ ﺑﻬﺪﺍﺷﺖ ﺟﻬﺎﻧﻲ )WHO( ﺍﺳﺘﻔﺎﺩﻩ ﺍﺯ ﻣﺎﺳﮏ ﻭ ﺭﻋﺎﻳﺖ ﻓﺎﺻﻠﻪ ﺍﺟﺘﻤﺎﻋﻲ ﺩﺭ ﻣﮑﺎﻥﻫﺎﻱ ﻋﻤﻮﻣﻲ ﻭ ﺷﻠﻮﻍ ﺭﺍ ﺑﻬﺘﺮﻳﻦ ﺭﻭﺵ ﭘﻴﺸﮕﻴﺮﺍﻧﻪ ﻣﻌﺮﻓﻲ ﮐﺮﺩﻩ ﺍﺳﺖ. ﺍﻳﻦ ﻣﻘﺎﻟﻪ ﻳﮏ ﺳﻴﺴﺘﻢ ﺑﺮﺍﻱ ﺷﻨﺎﺳﺎﻳﻲ ﺍﻓﺮﺍﺩ ﺩﺍﺭﺍﻱ ﻣﺎﺳﮏ ﭘﻴﺸﻨﻬﺎﺩ ﻣﻲﮐﻨﺪ ﮐﻪ ﺑﺮ ﭘﺎﻳﻪ ﻳﺎﺩﮔﻴﺮﻱ ﺍﻧﺘﻘﺎﻟﻲ ﻭ ﻣﻌﻤﺎﺭﻱ Inception v3 ﺍﺳﺖ. ﺭﻭﺵ ﭘﻴﺸﻨﻬﺎﺩﻱ ﺑﺎ ﺍﺳﺘﻔﺎﺩﻩ ﺍﺯ ﺩﻭ ﻣﺠﻤﻮﻋﻪ ﺩﺍﺩﻩ Simulated Mask Face Dataset) SMFD( ﻭ )MaskedFace-Net( MFN ﺁﻣﻮﺯﺵ ﻣﻲﺑﻴﻨﺪ ﻭ ﺑﺎ ﺗﻨﻈﻴﻢ ﺑﻬﻴﻨﻪ ﻓﺮﺍﭘﺎﺭﺍﻣﺘﺮﻫﺎ ﻭ ﻃﺮﺍﺣﻲ ﺩﻗﻴﻖ ﺑﺨﺶ ﺗﻤﺎﻣﺄ ﻣﺘﺼﻞ ﺳﻌﻲ ﻣﻲﮐﻨﺪ ﺩﻗﺖ ﺳﻴﺴﺘﻢ ﭘﻴﺸﻨﻬﺎﺩﻱ ﺭﺍ ﺍﻓﺰﺍﻳﺶ ﺩﻫﺪ. ﺍﺯ ﻣﺰﺍﻳﺎﻱ ﺳﻴﺴﺘﻢ ﭘﻴﺸﻨﻬﺎﺩﻱ ﺍﻳﻦ ﺍﺳﺖ ﮐﻪ ﻣﻲﺗﻮﺍﻧﺪ ﻋﻼﻭﻩ ﺑﺮ ﺻﻮﺭﺕﻫﺎﻱ ﺩﺍﺭﺍﻱ ﻣﺎﺳﮏ ﻭ ﺑﺪﻭﻥ ﻣﺎﺳﮏ، ﺣﺎﻟﺖﻫﺎﻱ ﺍﺳﺘﻔﺎﺩﻩ ﻏﻴﺮ ﺻﺤﻴﺢ ﺍﺯ ﻣﺎﺳﮏ ﺭﺍ ﻧﻴﺰ ﺗﺸﺨﻴﺺ ﺩﻫﺪ. ﺍﺯ ﺍﻳﻦﺭﻭ ﺭﻭﺵ ﭘﻴﺸﻨﻬﺎﺩﻱ ﺗﺼﺎﻭﻳﺮ ﭼﻬﺮﻩ ﻭﺭﻭﺩﻱ ﺭﺍ ﺑﻪ ﺳﻪ ﺩﺳﺘﻪ ﺗﻘﺴﻴﻢﺑﻨﺪﻱ ﺧﻮﺍﻫﺪ ﮐﺮﺩ. ﻧﺘﺎﻳﺞ ﺁﺯﻣﺎﻳﺸﻲ، ﺩﻗﺖ ﻭ ﮐﺎﺭﺍﻳﻲ ﺑﺎﻻﻱ ﺭﻭﺵ ﭘﻴﺸﻨﻬﺎﺩﻱ ﺭﺍ ﺩﺭ ﻣﻮﺿﻮﻉ ﻓﻮﻕ ﻧﺸﺎﻥ ﻣﻲﺩﻫﻨﺪ؛ ﺑﻄﻮﺭﻱﮐﻪ ﺍﻳﻦ ﻣﺪﻝ ﺩﺭ ﺩﺍﺩﻩﻫﺎﻱ ﺁﻣﻮﺯﺵ ﺑﻪ ﺩﻗﺖ 99.47 ﻭ ﺩﺭ ﺩﺍﺩﻩﻫﺎﻱ ﺁﺯﻣﺎﻳﺸﻲ ﺑﻪ ﺩﻗﺖ99.33 ﺩﺳﺖ ﻳﺎﻓﺘﻪ ﺍﺳﺖ.

Due to the epidemic of the coronavirus (Covid-19) and its rapid spread around the world, the world has faced a huge crisis. To prevent the spread of the coronavirus, the World Health Organization (WHO) has introduced the use of masks and keeping social distance as the best preventive method. So, developing an automatic monitoring system for detection of facemask in some crowded places is essential. To do this, we propose a mask recognition system based on transfer learning and Inception v3 architecture. In the proposed method, two datasets are used simultaneously for training including: Simulated Mask Face Dataset (SMFD) and MaskedFace-Net (MFN).this paper tries to increase the accuracy of the proposed system by optimally setting hyper-parameters and accurately designing the fully connected layers. The main advantage of the proposed method is that in addition to masked and unmasked face, it can also detect cases of incorrect use of mask. Therefore, the proposed method classifies the input face images into three categories. Experimental results show the high accuracy and efficiency of the proposed method; so that, this method has achieved to accuracy of 99.47% and 99.33% in training and test data respectively.