<p>Cancer is one of the most hazardous illnesses, which has been increasing the rate of mortality. In both men and women, lung cancer is recognized as the most widespread form of malignancy. The formation of unregulated cell growth in the lungs creates lung cancer. The earlier stage identification of lung cancer is complex, owing to its asymptomatic nature and long processing time. To solve these difficulties, a novel hybrid approach named Kronecker Mobile Forwarded Harmonic Network (KMFHNet)-based lung cancer detection is devised in this research. The proposed model integrates MobileNet for efficient spatial feature extraction, Deep Kronecker Network (DKN) for higher-order feature interaction, and forward harmonic analysis for enhanced frequency-domain representation. The CT lung image is provided as the input, and noise reduction is effectively achieved using the Sobel filter. Moreover, the lung lobes from the images are segmented with the utilization of the proposed Tversky- Dense-Res-Inception Net (DRI-Net), where the Tversky Similarity modifies the loss function. Image augmentation, such as sharpening, resizing, zooming, and shifting, is utilized for increasing the image size. After extracting the essential features, KMFHNet is employed to detect lung cancer. The effectiveness of the proposed model for lung cancer detection is validated by its high-performance metrics, including 92.55% accuracy, 91.38% recall, 90.89% precision, and an F1-score of 89.79%. The robustness and reliability of the KMFHNet framework for early-stage lung cancer identification are confirmed by these results.</p>