Internet of Medical Things: An Effective and Fully Automatic IoT Approach Using Deep Learning and Fine-Tuning to Lung...

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Internet of Things (IoT)

Background:

• Pulmonary diseases like COPD and tuberculosis pose significant global health challenges.
• Accurate detection of lung regions in medical images is crucial for diagnosing these pathologies.
• Computational methods, including IoT and deep learning, show promise in improving medical diagnosis.

Purpose of the Study:

• To propose a novel, automated model for the classification and segmentation of pulmonary CT images.
• To leverage Internet of Medical Things (IoMT) and deep learning with transfer learning for enhanced lung image analysis.
• To improve the accuracy and efficiency of lung disease diagnosis through advanced computational techniques.

Main Methods:

• Developed a new model integrating Internet of Things (IoT) for pulmonary CT image analysis.
• Applied transfer learning techniques within deep learning frameworks, combined with Parzen's probability density.
• Utilized an Application Programming Interface (API) based on the Internet of Medical Things for image classification and employed Mask R-CNN for segmentation.

Main Results:

• Achieved over 98% accuracy in the classification of pulmonary CT images.
• Reached a segmentation accuracy of 98.34% with a segmentation time of 5.43 seconds.
• The proposed fully automatic methodology outperformed existing transfer learning models in lung segmentation.

Conclusions:

• The proposed IoMT-based model offers a robust, fully automatic, and efficient solution for lung CT image classification and segmentation.
• This approach simplifies the segmentation process and provides superior performance compared to other methods.
• The model demonstrates significant potential for improving the early detection and diagnosis of pulmonary diseases.