LUNG CANCER PREDICTION USING IMPROVED VGG19 NETWORKS AND ADVANCED IMAGE PROCESSING TECHNIQUES

Lung cancer remains a leading cause of cancer-related deaths worldwide, necessitating advancements in early detection and accurate diagnosis. This study proposes a robust methodology for lung cancer prediction leveraging an improved VGG19 network in conjunction with advanced image processing techniques. The proposed framework integrates multiple stages to enhance the precision of detection and classification. Initially, noise in the lung imaging data is mitigated using a fuzzy-based median filter, which effectively reduces noise while preserving crucial structural details. This preprocessing step is critical for enhancing the quality of the input data and ensuring reliable feature extraction. Following noise removal, feature extraction is conducted using the Circular Local Binary Pattern (CLBP) method, CLBP captures the intricate textural and spatial information within the lung images, which is essential for distinguishing between benign and malignant regions. The segmented images are then processed through a threshold-based Mutilate Segmentation technique. This segmentation method accurately isolates the regions of interest (ROIs) within the lung images, ensuring that only relevant sections are analyzed in subsequent steps. Finally, the classification stage employs an improved VGG19 network, optimized for lung cancer detection. Modifications to the traditional VGG19 architecture enhance its capability to recognize patterns specific to lung cancer, resulting in higher accuracy and reduced false-positive rates. The integration of these advanced image processing techniques with the improved VGG19 network presents a comprehensive solution for lung cancer prediction. This study underscores the importance of combining sophisticated image processing methods with deep learning models to achieve superior performance in medical image analysis