Scholarlink

A medical practitioner employing computer intelligent systems proves beneficial for multiple investigations. Disease detection in recent times has significantly used machine learning, which opened new opportunities. Utilizing various diagnostic techniques such as medical imaging, biomarker analysis for accurately predicting the existence and severity of lung disorders is possible with artificial intelligence. Lung diseases such as symptoms of lung cancer are not apparent until the disease has advanced to a later stage. For this very reason it’s vital to analyze and predict pulmonary conditions in early stage to reduce burden on patient and healthcare system, and Machines learning helps in this regard which significantly improves the survival rate of a patient. The technique discussed in the article makes use of image segmentation method and convolutional neural network for feature extraction. The dataset used is the combined dataset of three open access datasets from kaggle, such as pneumonia dataset with 5856 images. The aim was early detection and diagnosis of presence of lung related diseases in Medical Images such as CXR images. We have created a web application for medical practitioner to analyze the images for presence of disease. The Accuracy increases as the number of epoch’s increases and accordingly loss decreases. In comparison to existing machine learning model which has accuracy between 94 to 98%, proposed model achieves higher detection accuracy of 99% using “EfficientNetB3”.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Detection of Lung Disease through Application of Machine Learning Algorithms with Transfer Learning

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Sagar Patil Baburao,
  • Suchita S. Patil,
  • Jyoti P. Kharade,
  • Shrikant Bhopale,
  • Shankar Madkar,
  • Rohit Raut

摘要

A medical practitioner employing computer intelligent systems proves beneficial for multiple investigations. Disease detection in recent times has significantly used machine learning, which opened new opportunities. Utilizing various diagnostic techniques such as medical imaging, biomarker analysis for accurately predicting the existence and severity of lung disorders is possible with artificial intelligence. Lung diseases such as symptoms of lung cancer are not apparent until the disease has advanced to a later stage. For this very reason it’s vital to analyze and predict pulmonary conditions in early stage to reduce burden on patient and healthcare system, and Machines learning helps in this regard which significantly improves the survival rate of a patient. The technique discussed in the article makes use of image segmentation method and convolutional neural network for feature extraction. The dataset used is the combined dataset of three open access datasets from kaggle, such as pneumonia dataset with 5856 images. The aim was early detection and diagnosis of presence of lung related diseases in Medical Images such as CXR images. We have created a web application for medical practitioner to analyze the images for presence of disease. The Accuracy increases as the number of epoch’s increases and accordingly loss decreases. In comparison to existing machine learning model which has accuracy between 94 to 98%, proposed model achieves higher detection accuracy of 99% using “EfficientNetB3”.