<p>Early identification of lung cancer plays a pivotal role in an individual’s survival, yet it poses a substantial challenge. Typically, Computed Tomography (CT) and Chest Radiographs (X-ray) scans serve as initial diagnostic tools for malignant nodules. However, there is a chance for incorrect diagnoses because of the presence of nodules. In the former stages, distinguishing the nodules of malignant and benign can be exceptionally challenging due to their striking similarities in appearance. Hence, this paper goal to develop a detection framework for lung nodules utilizing advanced deep structure techniques. The wanted lung nodule CT images are attained from several web sources, and the images are given to the phase of segmentation for getting lung nodule segmented images. Here, the Adaptive 3D trans-MobileUnet++ is utilized for segmenting the lung nodule images. In addition, the attributes presented within the Adaptive 3D trans-MobileUnet++ are tuned via the proposed COOT bird Optimization Strategy (COS) for enhancing the segmentation performance. Then, the images are forwarded to the detection stage, the Multi-scale Dilated MobilenetV2 with Attention Mechanism (MDMnet-AM) is used for detecting the lung nodules with higher effectiveness. The empirical analysis outcome is ensured with the conventional systems for validating the efficacy.</p>