Hendrik Leo, Khairun Saddami, Roslidar, R. Muharar, K. Munadi, F. Arnia
{"title":"基于cnn热图分类的肥胖症早期诊断移动应用","authors":"Hendrik Leo, Khairun Saddami, Roslidar, R. Muharar, K. Munadi, F. Arnia","doi":"10.1109/ICAIIC57133.2023.10066987","DOIUrl":null,"url":null,"abstract":"Obesity is one of the major risk factors for non-communicable diseases. Developing an early obese screening method is crucial to facilitate the early treatment of obese patients. In this study, we proposed a stand-alone mobile application for early diagnosis of obesity based on Convolution Neural Network (CNN) classifier model. The proposed CNN model was developed based on MobileNetV2 by modifying the fully connected layers. We trained the proposed model with the obese thermogram dataset through the transfer learning method and compared the classification performances with pre-trained models. The testing results show that the proposed model achieved an accuracy of 87.50%, a specificity of 100 %, and a sensitivity of 75.00 %. The proposed model demonstrated an optimal fit learning with 2.5 million learning parameters, a computation cost of 0.613 GFLOPs, and a size of 9.8 MB. The proposed model has been deployed and tested into the thermal camera smartphone CAT S62 Pro to do an early diagnosis of obesity.","PeriodicalId":105769,"journal":{"name":"2023 International Conference on Artificial Intelligence in Information and Communication (ICAIIC)","volume":null,"pages":null},"PeriodicalIF":0.0000,"publicationDate":"2023-02-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"A Mobile Application for Obesity Early Diagnosis Using CNN-based Thermogram Classification\",\"authors\":\"Hendrik Leo, Khairun Saddami, Roslidar, R. Muharar, K. Munadi, F. Arnia\",\"doi\":\"10.1109/ICAIIC57133.2023.10066987\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"Obesity is one of the major risk factors for non-communicable diseases. Developing an early obese screening method is crucial to facilitate the early treatment of obese patients. In this study, we proposed a stand-alone mobile application for early diagnosis of obesity based on Convolution Neural Network (CNN) classifier model. The proposed CNN model was developed based on MobileNetV2 by modifying the fully connected layers. We trained the proposed model with the obese thermogram dataset through the transfer learning method and compared the classification performances with pre-trained models. The testing results show that the proposed model achieved an accuracy of 87.50%, a specificity of 100 %, and a sensitivity of 75.00 %. The proposed model demonstrated an optimal fit learning with 2.5 million learning parameters, a computation cost of 0.613 GFLOPs, and a size of 9.8 MB. The proposed model has been deployed and tested into the thermal camera smartphone CAT S62 Pro to do an early diagnosis of obesity.\",\"PeriodicalId\":105769,\"journal\":{\"name\":\"2023 International Conference on Artificial Intelligence in Information and Communication (ICAIIC)\",\"volume\":null,\"pages\":null},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2023-02-20\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"2023 International Conference on Artificial Intelligence in Information and Communication (ICAIIC)\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1109/ICAIIC57133.2023.10066987\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"2023 International Conference on Artificial Intelligence in Information and Communication (ICAIIC)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/ICAIIC57133.2023.10066987","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
A Mobile Application for Obesity Early Diagnosis Using CNN-based Thermogram Classification
Obesity is one of the major risk factors for non-communicable diseases. Developing an early obese screening method is crucial to facilitate the early treatment of obese patients. In this study, we proposed a stand-alone mobile application for early diagnosis of obesity based on Convolution Neural Network (CNN) classifier model. The proposed CNN model was developed based on MobileNetV2 by modifying the fully connected layers. We trained the proposed model with the obese thermogram dataset through the transfer learning method and compared the classification performances with pre-trained models. The testing results show that the proposed model achieved an accuracy of 87.50%, a specificity of 100 %, and a sensitivity of 75.00 %. The proposed model demonstrated an optimal fit learning with 2.5 million learning parameters, a computation cost of 0.613 GFLOPs, and a size of 9.8 MB. The proposed model has been deployed and tested into the thermal camera smartphone CAT S62 Pro to do an early diagnosis of obesity.