Finding lung cancer frequently depends on analyzing small nodules in CT scans, which requires accurate segmentation techniques that go below basic recognition methods. Although current techniques that employ alternative architectures may attain respectable precision, they frequently encounter issues related to restricted CT scan databases and expansion, impeding their practical utility. By combining the Luna16 dataset against the Effective U-Net architecture, that is utilized to achieve superior performance on image classification problems whereas employing low computing power, our research addresses the urgent need for improved detection of lung cancer. This means that accurate results might be achieved about few parameters along with estimation is less expensive contrasting to other models. Our suggested model ensures sharper more precise nodule identification in CT scans by excelling in gathering features and conquering gradients that vanish. A varied and described standard, the Luna16 dataset, enables thorough learning and adaptation to different nodule kinds and imaging settings.