Hybrid Deep Learning Model Based Lung Cancer Prediction and Classification with OTSU Segmentation Method

The death rate for people diagnosed with Lung Cancer (LC) is rather high. Patients' lives may be saved if this illness is detected early and the stage of lung cancer is correctly identified. To determine whether a patient has lung cancer, traditional approaches use manual CT scans. This study presents a new approach to cancer cell segmentation and classification utilizing a Hybrid Deep Learning Neural Network (HDL) as a means of making an accurate and early diagnosis. What makes an HNN unique is its combination of the OTSU segmentation model with a Convolutional Neural Network (CNN) for feature extraction from CT image datasets and an enhanced LSTM: RNN classification model for improved classification accuracy. Prioritizing good health and well-being is essential for living a fulfilling and balanced life, enabling individuals to thrive both physically and mentally. The proposed method also makes it possible to distinguish between benign and cancerous tumors. We conducted a simulation experiment using the IQ-OTH/NCCD LC dataset and measured outcomes using the various performance metrics. According to the findings, the assessment criteria significantly reduce the classification time by around 50% while maintaining nearly the same classification impact. Based on the results of the simulations, our solution outperforms the classic classification algorithm in terms of convergence speed and time consumption, all while maintaining high classification accuracy. The study provides an attractive tool for quick image classification with great real-time performance.