Efficient Method to Classify Lung Cancer from CT Images Based on ARIMA-CF-LSTM Approach

Lung cancer is the most prevalent and lethal type of cancer, dramatically impacting global cancer-related mortality rates. Timely and precise detection is essential, and CT scans are regarded as the benchmark for screening due to their ability to furnish comprehensive details regarding tumor dimensions, positioning, and advancement. This study concentrates on the identification and examination of lung cancer through CT pictures utilizing a sophisticated deep learning framework. The methodology encompasses a series of steps: image preprocessing, segmentation, binarization, and feature extraction. Preprocessing activities, including image normalization, grayscale conversion, and noise elimination, prepare the CT scans for examination. Image binarization converts grayscale photos into monochrome, whereas segmentation partitions the image into significant components for targeted study. Feature extraction use a parametric statistical methodology to identify the most informative attributes. The suggested ARIMA-CF-LSTM model is subsequently trained using these features. The suggested hybrid model exhibited higher performance across all criteria when compared to two established techniques, ARIMA and LSTM. The method attained a remarkable accuracy of 96.25%, demonstrating its significant promise for facilitating the early detection and monitoring of lung cancer through CT imaging. This method presents a significant enhancement in medical image analysis for clinical use.