سنجش مزاج خشك و تر با استفاده از تصاوير حرارتي و طبقه بندي سلسله مراتبي

A hierarchical classification of dry/wet Mizaj (Temperament) using thermal imaging

ﺳﺎﺑﻘﻪ ﻭ ﻫﺪﻑ: ﺗﻌﻴﻴﻦ ﻣﻴﺰﺍﻥ ﺍﻧﺤﺮﺍﻑ ﺍﺯ ﻣﺰﺍﺝ ﭘﺎﻳﺔ ﻫﺮ ﻓﺮﺩ ﻳﻜﻲ ﺍﺯ ﻣﺮﺍﺣﻞ ﺍﻭﻟﻴﺔ ﺗﺸﺨﻴﺺ ﺑﻴﻤﺎﺭﻱ ﻭ ﺻﺤﺖ ﺍﻓﺮﺍﺩ ﺩﺭ ﻃﺐ ﺳﻨﺘﻲ ﺍﻳﺮﺍﻥ ﺍﺳﺖ. ﺍﻣﺎ ﺍﻳﻦ ﻣﺰﺍﺝ ﺳﻨﺠﻲ ﻭﺍﺑﺴﺘﮕﻲ ﺯﻳﺎﺩﻱ ﺑﻪ ﺑﻴﻤﺎﺭ ﻭ ﭘﺰﺷﻚ ﺩﺍﺭﺩ. ﻫﺪﻑ ﺍﺯ ﺍﻳﻦ ﺗﺤﻘﻴﻖ، ﻳﺎﻓﺘﻦ ﺭﻭﺵ ﻭ ﺍﺑﺰﺍﺭﻱ ﭘﺎﻳﺎ ﻭ ﺭﻭﺍ ﺑﻮﺩ ﻛﻪ ﺗﺎ ﺣﺪ ﺍﻣﻜﺎﻥ ﺑﺘﻮﺍﻧﺪ ﺳﻨﺠﺶ ﻣﺰﺍﺝ ﺭﺍ ﺍﺯ ﺣﺎﻟﺖ ﺫﻫﻨﻲ )ﺳﻮﺑﮋﻛﺘﻴﻮ( ﺑﻪ ﺣﺎﻟﺖ ﻋﻴﻨﻲ )ﺁﺑﮋﻛﺘﻴﻮ( ﺗﺒﺪﻳﻞ ﻛﻨﺪ. ﻣﻮﺍﺩ ﻭ ﺭﻭﺵ ﻫﺎ: ﺍﻳﻦ ﻣﻄﺎﻟﻌﻪ ﺩﺭ ﺯﻣﺴﺘﺎﻥ ﺳﺎﻝ 1398 ﺩﺭ ﺳﻼﻣﺘﻜﺪﺓ ﻃﺐ ﺳﻨﺘﻲ ﺑﻬﺸﺖ، ﻭﺍﺑﺴﺘﻪ ﺑﻪ ﺩﺍﻧﺸﮕﺎﻩ ﻋﻠﻮﻡ ﭘﺰﺷﻜﻲ ﺍﻳﺮﺍﻥ )ﺍﻳﺮﺍﻥ، ﺗﻬﺮﺍﻥ( ﺍﻧﺠﺎﻡ ﺷﺪ. ﺩﺭ ﺍﻳﻦ ﺗﺤﻘﻴﻖ 34 ﺩﺍﻭﻃﻠﺐ ﺍﺯ ﭘﺮﺳﻨﻞ ﭘﺲ ﺍﺯ ﺍﺧﺬ ﺭﺿﺎﻳﺖ ﻛﺘﺒﻲ ﺯﻳﺮ ﻧﻈﺮ ﻳﻚ ﭘﺰﺷﻚ ﻣﺘﺨﺼﺺ ﻃﺐ ﺍﻳﺮﺍﻧﻲ ﺑﺎ ﻓﺮﻡ ﺍﺳﺘﺎﻧﺪﺍﺭﺩﺷﺪﺓ ﻣﺠﺎﻫﺪﻱ، ﻣﺰﺍﺝ ﺳﻨﺠﻲ ﺷﺪﻧﺪ ﻭ ﻫﻢ ﺯﻣﺎﻥ ﺍﺯ ﻣﺤﻞ ﻟﻤﺲ ﻧﺒﺾ، ﭘﺸﺖ ﺩﺳﺖ ﻭ ﻛﻞ ﺻﻮﺭﺕ ﺑﺎ ﺍﺳﺘﻔﺎﺩﻩ ﺍﺯ ﺩﻭﺭﺑﻴﻦ ﺣﺮﺍﺭﺗﻲ ﻣﺪﻝ T2 ﺷﺮﻛﺖ ULIRVISION ﺗﺼﺎﻭﻳﺮ ﺣﺮﺍﺭﺗﻲ ﮔﺮﻓﺘﻪ ﻭ ﻣﻘﺪﺍﺭ ﺩﻣﺎﻱ ﻣﻮﺿﻊ ﺛﺒﺖ ﺷﺪ. ﻳﺎﻓﺘﻪ ﻫﺎ: ﺑﺎ ﺍﺳﺘﺨﺮﺍﺝ ﺩﻭﺍﺯﺩﻩ ﻭﻳﮋﮔﻲ ﻣﺮﺗﺒﻂ ﺑﺎ ﻧﺤﻮﺓ ﺗﻮﺯﻳﻊ ﺣﺮﺍﺭﺕ ﺩﺭ ﻧﺎﺣﻴﺔ ﻣﻮﺭﺩ ﻧﻈﺮ ﺩﺭ ﺗﺼﺎﻭﻳﺮ ﺣﺮﺍﺭﺗﻲ ﺫﻛﺮ ﺷﺪﻩ، ﻳﻚ ﺍﻟﮕﻮﺭﻳﺘﻢ ﻣﺒﺘﻨﻲ ﺑﺮ ﺍﻟﮕﻮﺭﻳﺘﻢ ژﻧﺘﻴﻚ ﻭ ﻃﺒﻘﻪ ﺑﻨﺪﻱ ﻛﻨﻨﺪﺓ Adaboost ﻃﺮﺍﺣﻲ ﺷﺪ ﻭ ﻣﺸﺎﻫﺪﻩ ﺷﺪ ﻛﻪ ﻫﻔﺖ ﻭﻳﮋﮔﻲ ﺍﺯ ﺍﻳﻦ ﻭﻳﮋﮔﻲ ﻫﺎ ﺑﻴﺸﺘﺮﻳﻦ ﺗﺄﺛﻴﺮ ﺭﺍ ﺩﺭ ﻃﺒﻘﻪ ﺑﻨﺪﻱ ﺩﺍﺭﻧﺪ ﻭ ﺻﺤﺖ 85/6 ﺩﺭﺻﺪ ﻭ ﺣﺴﺎﺳﻴﺖ 66/7 ﺩﺭﺻﺪ ﻭ ﻭﻳﮋﮔﻲ 91/7 ﺩﺭﺻﺪ ﺑﻪ ﺩﺳﺖ ﺁﻣﺪ. ﻧﺘﻴﺠﻪ ﮔﻴﺮﻱ: ﺩﺭ ﺍﻳﻦ ﻣﻘﺎﻟﻪ ﻧﺸﺎﻥ ﺩﺍﺩﻩ ﺷﺪ ﻛﻪ ﺗﺼﻮﻳﺮﺑﺮﺩﺍﺭﻱ ﺣﺮﺍﺭﺗﻲ ﺗﻄﺎﺑﻖ ﻣﻌﻨﻲ ﺩﺍﺭﻱ ﺑﺎ ﻓﺮﻡ ﺍﺳﺘﺎﻧﺪﺍﺭﺩﺷﺪﻩ ﺩﺍﺭﺩ ﻭ ﻣﻲ ﺗﻮﺍﻧﺪ ﺑﻪ ﻋﻨﻮﺍﻥ ﺍﺑﺰﺍﺭﻱ ﭘﺎﻳﺎ ﺑﺮﺍﻱ ﺍﺭﺯﻳﺎﺑﻲ ﻣﺰﺍﺝ خﺸﻚ/ ﺗﺮ ﻣﻮﺭﺩ ﺍﺳﺘﻔﺎﺩﻩ ﻗﺮﺍﺭ ﮔﻴﺮﺩ.

Background and Purpose: Mizaj (Temperament) identification is an important stage of diagnosis in Persian Medicine (PM). This study aimed to evaluate thermal imaging as a reliable tool that can be used instead of subjective assessments. Materials and Methods: This study was conducted in the winter of 2020 at the Behesht Traditional Medicine Health Clinic (Tehran, Iran). The Mizaj (Temperament) of 34 participants was assessed by a PM specialist using Mojahedi Mizaj standardized Questionnaire (MMQ). Also thermal images of the wrist at the radial pulse location, the back of the hand, and their whole face were recorded by a thermography camera model T2 (ULIRVISION Co.) and the temperatures of regions of interest were measured. Results: Twelve thermal features were extracted and a classifying algorithm was designed based on a genetic algorithm and an Adaboost classifier. By selecting 7 of the 12 features that had the most impact on the classification, the mean accuracy of 85.6%, sensitivity of 66.7%, and specificity of 91.7% were obtained. Conclusions: The thermal imaging was significantly consistent with MMQ and can be used as a reliable tool for evaluating dry/wet Mizaj (Temperament).