
Research Article
Improving Morphology and Recurrent Residual Refinement Network to Classify Hypertension in Retinal Vessel Image
@INPROCEEDINGS{10.1007/978-3-030-92942-8_2, author={Vo Thi Hong Tuyet and Nguyen Thanh Binh}, title={Improving Morphology and Recurrent Residual Refinement Network to Classify Hypertension in Retinal Vessel Image}, proceedings={Nature of Computation and Communication. 7th EAI International Conference, ICTCC 2021, Virtual Event, October 28--29, 2021, Proceedings}, proceedings_a={ICTCC}, year={2022}, month={1}, keywords={Morphology Saliency Recurrent residual refinement network Hypertension classification Segmentation}, doi={10.1007/978-3-030-92942-8_2} }
- Vo Thi Hong Tuyet
Nguyen Thanh Binh
Year: 2022
Improving Morphology and Recurrent Residual Refinement Network to Classify Hypertension in Retinal Vessel Image
ICTCC
Springer
DOI: 10.1007/978-3-030-92942-8_2
Abstract
Prolonged or severe hypertension leads to vascular changes, resulting in endothelial damage and necrosis. The hypertension classification based on the retinal blood vessel segmentation is one of the state-of-the-art approaches. Therefore, the detection and classification of hypertensive retinal images is very useful in the diagnosis of hypertension. This paper proposed a method for the hypertension classification based on saliency, in which the extracting discriminative features keep information in the spatial domain. It includes three stages: morphology for enhancing the quality of the parameters, recurrent residual refinement network for feature extractions of the saliency and hypertension classification. The result of the proposed method is about 92.9% and better results than others in the STARE and DRIVE dataset.