Conditional Generative Adversarial Networks with Multi-scale Discriminators for Prostate MRI Segmentation
- PDF / 3,983,976 Bytes
- 11 Pages / 439.37 x 666.142 pts Page_size
- 44 Downloads / 204 Views
Conditional Generative Adversarial Networks with Multi-scale Discriminators for Prostate MRI Segmentation Jun He1 · Xinke Li1 · Ninghui Liu2 · Shu Zhan1
© Springer Science+Business Media, LLC, part of Springer Nature 2020
Abstract Accurate prostate MR image segmentation is a necessary preprocessing stage for computerassisted diagnostic algorithms. Convolutional neural network, as a research focus in recent years, has been proven to be powerful in computer vision field. Recently, the most effective prostate MRI segmentation technology mainly relies on full convolutional network which has been widely used in semantic segmentation task. However, it’s independent and identically distributed assumption neglect the structural regularity present in MR images and miss information between pixels. In this paper, we propose an MRI-conditional generative adversarial networks for prostate segmentation. Our adversarial training make it context aware and the use of adversarial loss functions learn high-level structural information. The network consist of a generator and a discriminator. The generator consists of a contraction channel and an expansion channel like U-Net. The method we proposed uses a multi-scale discriminator which consist of two discriminators with the same structure but different input sizes. The objective function has two parts: one is the adversarial loss, the other is feature matching loss which stabilizes the training and get better convergence. The experiment show that our network can accurately segment the prostate MRI and outperforms most existing methods. Keywords Magnetic resonance images (MRI) · Generative adversarial networks · Generator · Feature matching loss
1 Introduction Magnetic resonance imaging is widely used in clinical applications such as tumor detection because it can acquire clear images with high contrast. For the detection and treatment planning of prostate lesions, doctor first needs to outline the prostate tissue, differentiating between the prostate and surrounding organs, and then decide to take the appropriate treat-
B
Shu Zhan [email protected] Jun He [email protected]
1
School of Computer and Information, Hefei University of Technology, Hefei 230009, China
2
The First Affiliated Hospital of Anhui Medical University, Hefei 230601, China
123
J. He et al.
Fig. 1 Different prostate MRI of different people displaying large variations
ment. But this process currently relies mainly on manual segmentation by doctors. This is a very time consuming process, and the results of segmentation vary from person to person. Prostate MRI segmentation remains a challenging task for the below reason. (1) The contrast between the prostate tissue and other surrounding tissues is low, their boundaries are difficult to distinguish, as shown in Fig. 1. (2) The area belonging to the prostate tissue in a magnetic resonance image is small, and the available effective information is less. (3) The longer algorithm time consumption may delay clinical diagnosis. In order to solve the above probl
Data Loading...
