Hybrid healthcare unit recommendation system using computational techniques with lung cancer segmentation

Area of Science:

• Medical Imaging Analysis
• Artificial Intelligence in Healthcare
• Radiology

Background:

• Accurate lung nodule segmentation in Computed Tomography (CT) is crucial for diagnosis and treatment planning.
• Determining the particle composition of lung nodules is a vital aspect of patient care.

Purpose of the Study:

• To develop and evaluate a novel segmentation model for identifying lung disease from CT scans.
• To improve the accuracy of lung nodule detection and characterization.

Main Methods:

• A hybrid deep learning architecture, U-Net++, combining U-Net with a Two-parameter logistic distribution was proposed.
• Contrast Limited Adaptive Histogram Equalization (CLAHE) was applied to a dataset of 5,000 CT scan images.
• The model was trained and evaluated on the LUNA-16 dataset using various deep learning classifiers.

Main Results:

• The proposed U-Net++ model demonstrated superior performance in segmentation metrics including Probabilistic Rand Index (PRI), Variation of Information (VOI), Region of Interest (ROI), Dice Coefficient, and Global Consistency Error (GCE).
• Parameter estimation achieved a high accuracy of 91.76%, validating the model's effectiveness.

Conclusions:

• The U-Net++ model offers a significant advancement in lung nodule segmentation accuracy for CT images.
• This enhanced segmentation capability supports more precise diagnosis and personalized treatment planning for lung diseases.