A Hybrid Framework for Lung Cancer Classification

Cancer is the second leading cause of death worldwide, and rate lung cancer much higher than other types cancers. In recent years, numerous novel computer-aided diagnostic techniques with deep learning have been designed to detect in early stages. However, models are easy overfit, overfitting problem always causes lower performance. To solve this classification tasks, we proposed a hybrid framework called LCGANT. Specifically, our contains two main parts. The first part convolutional GAN (LCGAN) generate synthetic images. regularization enhanced transfer model VGG-DF classify images into three classes. Our achieves result 99.84%±0.156% (accuracy), 99.84%±0.153% (precision), (sensitivity), (F1-score). reaches highest performance dataset for task. resolves it better state-of-the-art methods.