Segmentation of CT Lung Images Using FCM with Active Contour and CNN Classifier.

Area of Science:

• Medical Imaging
• Artificial Intelligence in Medicine
• Oncology

Background:

• Lung cancer is a leading cause of mortality, often diagnosed at advanced stages.
• Early detection of lung tumors is challenging, requiring expert radiologist interpretation.
• Accurate segmentation of tumoral tissue in CT scans is crucial for timely diagnosis.

Purpose of the Study:

• To develop an effective method for segmenting tumoral lung tissue from CT images.
• To improve the early detection of lung cancer.
• To aid radiologists in identifying cancerous regions.

Main Methods:

• A hybrid segmentation technique combining active contour and Fuzzy c-means was employed.
• The segmented regions were classified as normal or abnormal using a Convolutional Neural Network (CNN).
• Image preprocessing, binarization, thresholding, and Gray-Level Co-occurrence Matrix (GLCM) feature extraction were utilized.

Main Results:

• The proposed hybrid segmentation method demonstrated high accuracy in segmenting lung tumors.
• Quantitative evaluation showed good accuracy (96.67%), Peak Signal-to-Noise Ratio (PSNR), and Mean Square Error (MSE) values.
• The technique effectively differentiates between normal and abnormal lung tissue.

Conclusions:

• The hybrid segmentation technique provides an effective approach for lung tumor segmentation in CT images.
• The combination of active contour, Fuzzy c-means, and CNN shows promise for improving lung cancer diagnosis.
• Further research can explore additional techniques to enhance segmentation details for future improvements.