LAC-GAN: Lesion attention conditional GAN for Ultra-widefield image synthesis. (January 2023)

• Record Type: Journal Article
• Authors: Lei, Haijun; Tian, Zhihui; Xie, Hai; Zhao, Benjian; Zeng, Xianlu; Cao, Jiuwen; Liu, Weixin; Wang, Jiantao; Zhang, Guoming; Wang, Shuqiang; Lei, Baiying
• Is Part Of: Neural networks. Volume 158(2023)
• Page Start: 89
• Page End: 98
• Publication Date: 2023-01
• Subjects: Ultra-widefield image -- Disease detection -- Conditional GAN -- Lesion attention; Neural computers -- Periodicals; Neural networks (Computer science) -- Periodicals; Neural networks (Neurobiology) -- Periodicals; Nervous System -- Periodicals; Ordinateurs neuronaux -- Périodiques; Réseaux neuronaux (Informatique) -- Périodiques; Réseaux neuronaux (Neurobiologie) -- Périodiques; Neural computers; Neural networks (Computer science); Neural networks (Neurobiology); Periodicals; 006.32
• DOI: 10.1016/j.neunet.2022.11.005
• Languages: English
• ISSNs: 0893-6080

Abstract

Abstract: Automatic detection of retinal diseases based on deep learning technology and Ultra-widefield (UWF) images plays an important role in clinical practices in recent years. However, due to small lesions and limited data samples, it is not easy to train a detection-accurate model with strong generalization ability. In this paper, we propose a lesion attention conditional generative adversarial network (LAC-GAN) to synthesize retinal images with realistic lesion details to improve the training of the disease detection model. Specifically, the generator takes the vessel mask and class label as the conditional inputs, and processes the random Gaussian noise by a series of residual block to generate the synthetic images. To focus on pathological information, we propose a lesion feature attention mechanism based on random forest (RF) method, which constructs its reverse activation network to activate the lesion features. For discriminator, a weight-sharing multi-discriminator is designed to improve the performance of model by affine transformations. Experimental results on multi-center UWF image datasets demonstrate that the proposed method can generate retinal images with reasonable details, which helps to enhance the performance of the disease detection model.