{"title":"基于大规模训练自组织映射和学习向量量化的肺结节分类","authors":"Yan Soon Weei, H. S. Pheng","doi":"10.1109/AiDAS47888.2019.8970882","DOIUrl":null,"url":null,"abstract":"The abnormal growth of cells in the lungs leads to the development of nodules and the overgrowth of lung nodules will eventually form a cancerous cell. Detection of lung nodules in the early stage is vital in such a way that proper treatments can be applied before the lung nodules grow into lethal lung cancer. In recent decades, machine learning has been widely used in the computer aided system to provide second opinion to the radiologists in the detection of abnormality on medical images. The aim of this paper is to implement a machine learning algorithm in the classification and enhancement of lung nodules on computed tomography (CT) images. The classification model – Massive-Training Self-Organizing Map and Learning Vector Quantization (MTSOM-LVQ) is implemented to classify the sub-regions based on the teaching Gaussian values. Each sub-region is associated with its teaching value generated by using Gaussian distribution function. The results show that MTSOM-LVQ is able to enhance nodules and suppressing non-nodules on CT images. Adjustment on the parameters such as map size, training iteration and size of the training sample would affect the performance of the MTSOMLVQ. Besides, the performance of the MTSOM-LVQ is validated and 90% classification sensitivity is achieved. As a conclusion, the training accuracy can be further improved by choosing the optimized parameters for MTSOM-LVQ in future research.","PeriodicalId":227508,"journal":{"name":"2019 1st International Conference on Artificial Intelligence and Data Sciences (AiDAS)","volume":null,"pages":null},"PeriodicalIF":0.0000,"publicationDate":"2019-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Lung Nodules Classification Using Massive-Training Self-Organizing Map and Learning Vector Quantization\",\"authors\":\"Yan Soon Weei, H. S. Pheng\",\"doi\":\"10.1109/AiDAS47888.2019.8970882\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"The abnormal growth of cells in the lungs leads to the development of nodules and the overgrowth of lung nodules will eventually form a cancerous cell. Detection of lung nodules in the early stage is vital in such a way that proper treatments can be applied before the lung nodules grow into lethal lung cancer. In recent decades, machine learning has been widely used in the computer aided system to provide second opinion to the radiologists in the detection of abnormality on medical images. The aim of this paper is to implement a machine learning algorithm in the classification and enhancement of lung nodules on computed tomography (CT) images. The classification model – Massive-Training Self-Organizing Map and Learning Vector Quantization (MTSOM-LVQ) is implemented to classify the sub-regions based on the teaching Gaussian values. Each sub-region is associated with its teaching value generated by using Gaussian distribution function. The results show that MTSOM-LVQ is able to enhance nodules and suppressing non-nodules on CT images. Adjustment on the parameters such as map size, training iteration and size of the training sample would affect the performance of the MTSOMLVQ. Besides, the performance of the MTSOM-LVQ is validated and 90% classification sensitivity is achieved. As a conclusion, the training accuracy can be further improved by choosing the optimized parameters for MTSOM-LVQ in future research.\",\"PeriodicalId\":227508,\"journal\":{\"name\":\"2019 1st International Conference on Artificial Intelligence and Data Sciences (AiDAS)\",\"volume\":null,\"pages\":null},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2019-09-01\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"2019 1st International Conference on Artificial Intelligence and Data Sciences (AiDAS)\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1109/AiDAS47888.2019.8970882\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"2019 1st International Conference on Artificial Intelligence and Data Sciences (AiDAS)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/AiDAS47888.2019.8970882","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
Lung Nodules Classification Using Massive-Training Self-Organizing Map and Learning Vector Quantization
The abnormal growth of cells in the lungs leads to the development of nodules and the overgrowth of lung nodules will eventually form a cancerous cell. Detection of lung nodules in the early stage is vital in such a way that proper treatments can be applied before the lung nodules grow into lethal lung cancer. In recent decades, machine learning has been widely used in the computer aided system to provide second opinion to the radiologists in the detection of abnormality on medical images. The aim of this paper is to implement a machine learning algorithm in the classification and enhancement of lung nodules on computed tomography (CT) images. The classification model – Massive-Training Self-Organizing Map and Learning Vector Quantization (MTSOM-LVQ) is implemented to classify the sub-regions based on the teaching Gaussian values. Each sub-region is associated with its teaching value generated by using Gaussian distribution function. The results show that MTSOM-LVQ is able to enhance nodules and suppressing non-nodules on CT images. Adjustment on the parameters such as map size, training iteration and size of the training sample would affect the performance of the MTSOMLVQ. Besides, the performance of the MTSOM-LVQ is validated and 90% classification sensitivity is achieved. As a conclusion, the training accuracy can be further improved by choosing the optimized parameters for MTSOM-LVQ in future research.