AUTHOR=Wang Tingting , Wang Meng , Zhu Weifang , Wang Lianyu , Chen Zhongyue , Peng Yuanyuan , Shi Fei , Zhou Yi , Yao Chenpu , Chen Xinjian 

TITLE=Semi-MsST-GAN: A Semi-Supervised Segmentation Method for Corneal Ulcer Segmentation in Slit-Lamp Images

JOURNAL=Frontiers in Neuroscience

VOLUME=Volume 15 - 2021

YEAR=2022

URL=https://www.frontiersin.org/journals/neuroscience/articles/10.3389/fnins.2021.793377

DOI=10.3389/fnins.2021.793377

ISSN=1662-453X

ABSTRACT=Corneal ulcer is a common leading cause illness of corneal blindness. It is difficult to accurately segment corneal ulcers due to the following problems: large difference in the pathological shapes between point-flaky and flaky corneal ulcers, burred boundary, noise interference and the lack of sufficient slit lamp images with ground truth. To address these problems, in this paper, we propose a novel semi-supervised multi-scale self-transformer generative adversarial network (Semi-MsST-GAN), which can leverage unlabeled images to improve the performance of corneal ulcer segmentation in fluorescein staining slit lamp images. First, to improve the performance of segmenting the corneal ulcer regions with complex pathological features, we propose a novel multi-scale self-transformer network (MsSTNet) as MsST-GAN 's generator, which can guide the model to aggregate the low-level weak semantic features with high-level strong semantic information and adaptively learn the spatial correlation in feature maps. Then, to further improve the segmentation performance by leveraging unlabeled data, the semi-supervised approach based on the proposed MsST-GAN is explored to solve the problem that lacks slit lamp images with corresponding ground truth. The proposed Semi-MsST-GAN was comprehensively evaluated on the public SUSTech-SYSU dataset which contains 354 labeled and 358 unlabeled fluorescein staining slit-lamp images. The results show that compared with other state-of-the-art methods, our proposed method achieves better performance with comparable efficiency.