Three-Dimensional Liver Image Segmentation Using Generative Adversarial Networks Based on Feature Restoration.

Runnan He,Henggui Zhang,Shiqi Xu,Na Zhao,Qince Li,Yang Liu,Yongfeng Yuan,Yashu Liu

doi:10.3389/fmed.2021.794969

Abstract

Medical imaging provides a powerful tool for medical diagnosis. In the process of computer-aided diagnosis and treatment of liver cancer based on medical imaging, accurate segmentation of liver region from abdominal CT images is an important step. However, due to defects of liver tissue and limitations of CT imaging procession, the gray level of liver region in CT image is heterogeneous, and the boundary between the liver and those of adjacent tissues and organs is blurred, which makes the liver segmentation an extremely difficult task. In this study, aiming at solving the problem of low segmentation accuracy of the original 3D U-Net network, an improved network based on the three-dimensional (3D) U-Net, is proposed. Moreover, in order to solve the problem of insufficient training data caused by the difficulty of acquiring labeled 3D data, an improved 3D U-Net network is embedded into the framework of generative adversarial networks (GAN), which establishes a semi-supervised 3D liver segmentation optimization algorithm. Finally, considering the problem of poor quality of 3D abdominal fake images generated by utilizing random noise as input, deep convolutional neural networks (DCNN) based on feature restoration method is designed to generate more realistic fake images. By testing the proposed algorithm on the LiTS-2017 and KiTS19 dataset, experimental results show that the proposed semi-supervised 3D liver segmentation method can greatly improve the segmentation performance of liver, with a Dice score of 0.9424 outperforming other methods.

Highlights

Recent advances in deep convolutional neural networks (DCNN) have shown great promises in handling many computer vision tasks such as target detection, image classification, and semantic segmentation, which can usually reach human-level performance
This paper proposed a semi-supervised 3D liver segmentation method based on deep convolutional generative adversarial networks (GAN) (DCGAN), which consists of the discriminator and generator
The experiment proves that the fake image is very close to the real image by the feature restoration method, and the speed of the network learning distribution is faster than that of the network using random noise as the input, which avoids the problem of gradient dispersion

Summary

Introduction

Recent advances in deep convolutional neural networks (DCNN) have shown great promises in handling many computer vision tasks such as target detection, image classification, and semantic segmentation, which can usually reach human-level performance. One of the main limitations of DCNN is that they require a large amount of labeled data for training process. This limitation is prominent in dealing with medical image segmentation problems. The acquisition of labeled three-dimensional (3D) medical images requires manual annotation, which is time-consuming and labor-intensive, limiting the further development of DCNN in medical image processing. The deep neural network needs a large dataset to train the model for obtaining the model parameters. In the absence of sufficient training data, the neural network will have relatively low performance and poor generalization ability

Methods

Results

Conclusion