Publications

You can also find my articles on my Google Scholar profile.

Journal Articles

Parallel Multi-Path Network for Ocular Disease Detection Inspired by Visual Cognition Mechanism

Published in IEEE Journal of Biomedical and Health Informatics, 2024

Various ocular diseases such as cataracts, glaucoma, and diabetic retinopathy have become several major factors causing non-congenital visual impairment, which seriously threatens vision health. The shortage of ophthalmic medical resources has brought huge obstacles to large-scale ocular disease screening. Therefore, it is necessary to use computer-aided diagnosis (CAD) technology to achieve large-scale screening and diagnosis of ocular diseases. In this work, inspired by the human visual cognition mechanism, we propose a parallel multi-path network for multiple ocular diseases detection, called PMP-OD, which integrates the detection of multiple common ocular diseases, including cataracts, glaucoma, diabetic retinopathy, and pathological myopia. The bottom-up features of the fundus image are extracted by a common convolutional module, the Low-level Feature Extraction module, which simulates the non-selective pathway. Simultaneously, the top-down vessel and other lesion features are extracted by the High-level Feature Extraction module that simulates the selective pathway. The retinal vessel and lesion features can be regarded as task-driven high-level semantic information in the physician disease diagnosis process. Then, the features are fused by a feature fusion module based on the attention mechanism. Finally, the disease classifier gives prediction results according to the integrated multi-features. The experimental results indicate that our PMP-OD model outperforms other state-of-the-art (SOTA) models on an ocular disease dataset reconstructed from ODIR-5K, APTOS-2019, ORIGA-light, and Kaggle.

Recommended citation: T. Deng, Y. Huang and C. Yang, "Parallel Multi-Path Network for Ocular Disease Detection Inspired by Visual Cognition Mechanism," in IEEE Journal of Biomedical and Health Informatics, vol. 29, no. 1, pp. 345-357, Jan. 2025.
Download Paper

Multi-level spatial-temporal and attentional information deep fusion network for retinal vessel segmentation

Published in Physics in Medicine and Biology, 2023

Automatic segmentation of fundus vessels has the potential to enhance the judgment ability of intelligent disease diagnosis systems. Even though various methods have been proposed, it is still a demanding task to accurately segment the fundus vessels. The purpose of our study is to develop a robust and effective method to segment the vessels in human color retinal fundus images. We present a novel multi-level spatial-temporal and attentional information deep fusion network for the segmentation of retinal vessels, called MSAFNet, which enhances segmentation performance and robustness. Our method utilizes the multi-level spatial-temporal encoding module to obtain spatial-temporal information and the Self-Attention module to capture feature correlations in different levels of our network. Based on the encoder and decoder structure, we combine these features to get the final segmentation results. Through abundant experiments on four public datasets, our method achieves preferable performance compared with other SOTA retinal vessel segmentation methods. Our Accuracy and Area Under Curve achieve the highest scores of 96.96%, 96.57%, 96.48% and 98.78%, 98.54%, 98.27% on DRIVE, CHASE_DB1, and HRF datasets. Our Specificity achieves the highest score of 98.58% and 99.08% on DRIVE and STARE datasets. The experimental results demonstrate that our method has strong learning and representation capabilities and can accurately detect retinal blood vessels, thereby serving as a potential tool for assisting in diagnosis.

Recommended citation: Y. Huang and T. Deng. (2023). "Multi-level spatial-temporal and attentional information deep fusion network for retinal vessel segmentation." Physics in Medicine and Biology. 68(195026).
Download Paper

Conference Papers

STSANet: Retinal Vessel Segmentation via Spatial-Temporal and Self-Attention Encoding

Published in 2022 14th International Conference on Wireless Communications and Signal Processing (WCSP), 2022

Retinal vessel segmentation can improve the judgment ability of intelligent disease diagnosis system. Although a large number of retinal vessel segmentation models have been proposed with the development of deep learning, it is still a challenging task. In this work, we propose a new retinal vessel segmentation network via spatial-temporal and self-attention encoding modules, called STSANet, which can significantly improve the performance and robustness of segmentation. The spatial-temporal information of fundus images are extracted by a Spatial-Temporal encoding module in the STSANet. In addition, the internal correlation of features is captured by the Self-Attention module. By fusing spatial-temporal and self-attention features, the final result contains both spatial-temporal information and internal feature information of fundus images. The experimental results indicate that our STSANet outperforms other state-of-the-art retinal segmentation models on the published standard datasets.

Recommended citation: Y. Huang and T. Deng. (2022). "STSANet: Retinal Vessel Segmentation via Spatial-Temporal and Self-Attention Encoding." 2022 14th International Conference on Wireless Communications and Signal Processing (BIC-TA). pp. 132-137.
Download Paper

MLFF: Multiple Low-Level Features Fusion Model for Retinal Vessel Segmentation

Published in The 16th International Conference on Bio-inspired Computing: Theories and Applications (BIC-TA), 2021

Imaging is increasingly used for the diagnosis of retinal normality and the monitoring of retinal abnormalities. Many retinal vessel properties, such as small artery aneurysms, narrowing of incisions, etc., are related to systemic diseases. The morphology of retinal blood vessels themselves is related to cardiovascular disease and coronary artery disease in adults. The fundus image can intuitively reflect the retinal vessel lesions, and the computer-based image processing method can be used for auxiliary medical diagnosis. In this paper, a retinal vessel segmentation model, named as MLFF, is proposed to effectively extract and fuse multiple low-level features. Firstly, there are 25 low-level feature maps of fundus retinal vessel images that are analyzed and extracted. Then, the feature maps are fused by an AdaBoost classifier. Finally, the MLFF is trained and evaluated on public fundus images for vessel extraction dataset (DRIVE). The qualitative and quantitative experimental results show that our model can effectively detect the retinal vessels and outperforms other models including deep learning-based models.

Recommended citation: Deng, T., Huang, Y., Zhang, J. (2022). "MLFF: Multiple Low-Level Features Fusion Model for Retinal Vessel Segmentation." 2021 International Conference on Bio-Inspired Computing: Theories and Applications (BIC-TA). pp. 271-281.
Download Paper

Yi Huang