Multie-scale hierarchical feature extraction combined with dilated-residual U-Net for retina automatic segmentation

Journal Title: Eye Science (Yanke Xuebao) - Year 2021, Vol 36, Issue 1

Abstract

[Objective:] To achieve the segmentation of different layers and fluid areas on the optical coherence tomography (OCT) image of the retina. [Methods:] A lightweight neural network based on deep learning was proposed. The network structure adopted in this study was designed based on the architecture of dilated-residual U-Net. By connecting the upsampling output obtained at different depth networks, multi-scale feature fusion was performed to enable the system to accurately identify the boundaries on the OCT image. [Results:] Compared with U-Net, this algorithm could achieve the same accuracy with 1–2 epochs less, and the accuracy was also improved by 1.25%. [Conclusion:] The proposed network improves the segmentation performance of retinal OCT images, and reduces the number of parameters, which demonstrates the network has great application potential.

Authors and Affiliations

Shanshan LIANG, Hongwei ZENG, Jie HE, Jun ZHANG, Jin YUAN

Keywords

Related Articles

Clinical evaluation of modified "Z"-shaped knotless transscleral suture fixation posterior chamber intraocular lens surgery for Congenital ectopia lentis

[Objective:] To evaluate the clinical efficacy of modified ā€œZā€-shaped knotless transscleral suture fixation intraocular lens (IOL) and traditional knotted transscleral suture fixation IOL in congenital ectopia lentis. [M...

Research progress in the treatment of chronic ocular graft-versus-host disease

More than half of the patients developed chronic graft-versus-host disease after accepting allogeneic hematopoietic stem cell transplantation suffer from ocular graft-versus-host disease. Ocular graft-versus-host disease...

Application of artificial intelligence in ocular fundus diseases

Artificial intelligence (AI) is about simulating and expanding human intelligence. AI based on deep learning (DL) can analyze images well by using their inherent features, such as outlines, frames and so on. As researche...

Congenital optic disc pit with macular retinoschisis: A case report

Optic disc pit is a rare congenital anomaly of the optic disc. Optic disc pit maculopathy, including macular retinoschisis and macular detachment, is the important cause of vision loss in patients. The origin of the subr...

Intervention plan for visual field defects in patients with stroke: a scoping review

[Objective:] To systematically analyze the related studies on intervention schemes for patients with visual field defects after stroke, and to identify, summarize, and summarize the specific content, outcome indicators,...

Download PDF file
  • EP ID EP763458
  • DOI 10.3978/j.issn.1000-4432.2021.01.10
  • Views 25
  • Downloads 0

How To Cite

Shanshan LIANG, Hongwei ZENG, Jie HE, Jun ZHANG, Jin YUAN (2021). Multie-scale hierarchical feature extraction combined with dilated-residual U-Net for retina automatic segmentation. Eye Science (Yanke Xuebao), 36(1), -. https://europub.co.uk/articles/-A-763458