<p>Lung cancer (LC) is one of the most severe and deadly diseases globally, with high mortality rates affecting both men and women. Improving patient outcomes requires an early and precise diagnosis, and Computed Tomography (CT) scans are a vital tool for identifying lung problems. To improve diagnosis accuracy, these images are increasingly being used in conjunction with machine learning (ML) and deep learning (DL) approaches. However, traditional deep neural networks often function as “black boxes” and suffer from complex, parameter-sensitive training processes that can impact their stability and effectiveness. To overcome these challenges, this research introduces a novel Enhanced Inception Layer-based GoogleNet with Transfer Learning (EIGN-TL) model for lung cancer classification. This proposed approach involves four stages, including preprocessing, segmentation, feature extraction, and classification. Initially, the resizing and standardization methods are applied to ensure consistent image dimensions and improve training stability. After preprocessing, the image is given to the segmentation phase, where the Improved Profuse Clustering Technique (IPCT) is proposed to enhance the segmentation accuracy. Subsequently, appropriate features such as Modified Quantize Orientation-based Local Gradient Increasing Pattern (MQO-LGIP), Pyramid Histogram of Oriented Gradients (PHOG), morphological features, and shape features are extracted from the segmented outcome to effectively capture the local gradient variations and texture changes crucial for accurate classification. Finally, the obtained feature set is given to the Enhanced Inception layer-based GoogleNet with Transfer Learning (EIGN-TL) model and produces the final classified outcomes. Moreover, the experimental results show that the proposed model achieves superior results compared to existing methods, with an accuracy of 95.7% and an F-measure of 0.955, demonstrating its robustness and reliability in the classification process.</p>