Deep Learning-based CT Image Segmentation for Liver Cancer

According to the global cancer statistics released in 2020, liver cancer is recognized as the sixth most common type of cancer globally and ranks third among the causes of death due to cancer. The high mortality rate of this cancer poses a great threat and serious harm to human life and health. Typically, liver cancer is diagnosed through liver biopsy or by utilizing enhanced computed tomography (CT) and magnetic resonance imaging (MRI) techniques. However, traditional diagnostic methods often rely on the physician's visual observation, which is not only time-consuming and very exhausting, but also subject to large errors and instability. To address these challenges, this paper provides an in-depth discussion of deep learning-based techniques for segmenting the liver and tumor in abdominal CT images. We propose a CT image segmentation method for liver cancer based on U-Net+SENet network, aiming to take advantage of the high accuracy and high training efficiency of U-Net+SENet network in image segmentation in order to assist imaging physicians to make more accurate and faster judgments. With this method, we hope to significantly improve the accuracy and efficiency of liver cancer diagnosis, and thus provide better treatment options and prognosis assessment for patients.