Review on Lung Cancer Lesion Detection and Segmentation Methods

Abstract

Lung cancer itself and relevant detection and segmentation methods, in the modern society, becomes increasingly popular and significant topics. Scientists believe that people smoke positively may deteriorate their body health themselves, and people who breathe it in second hand may also suffer from this harmful environment. To help people with lung cancer lesions, there are several methods used for cancer treatment. Automated CT imaging can encircle suggested segmentation areas in a 3-D version, and it provides convenience with users when they feel tired after diagnosing for a whole day. Semi-automated CT deep learning model is another technique to detect particular regions in the lung by adjusting pixels. Additionally, few-shot learning based on advanced learning algorithm is an efficient method for lung cancer lesion detection, and Generative Adversarial Networks(GAN) can be used for lung cancer detection by using a small number of medical images as train datasets. However, CNN model cannot obtain global information; therefore, the integration of 2dcnn and 3dcnn solves this limitation in an effective approach.