Lung Nodule Segmentation and Classification using U-Net and Efficient-Net

The ability to detect lung cancer has led better health outcomes. Deep learning techniques are widely used in the medical field tumors at an early stage. models such as U-Net, Efficient-Net, Resnet, VGG-16, etc. have been incorporated various studies accurately. To enhance detection performance, this work proposes algorithm that combines U-Net and Efficient-Net neural networks for nodule segmentation classification. A feature-extraction-based semi-supervised method is take advantage of huge amount CT scan images with no pathological labels. Semi-supervised achieved using a feature pyramid network (FPN) ResNet-50 model extraction classifier predicting unlabelled nodules. main innovation skip-connections, which give decoder access features encoder learned scales enable accurate localization uses depth, width, resolution scaling, combined compound coefficient uniformly all dimensions, resulting efficient image This evaluated on publicly available LIDC-IDRI dataset outperforms most existing methods. proposed aims address issues high false-positive rate, small nodules, wide range non-uniform longitudinal data. Experiment results show higher accuracy 91.67% when compared previous works.