Layer segmentation method and system for retina layer and effusion area based on deep learning

A retinal layer, deep learning technology, applied in neural learning methods, image analysis, biological neural network models, etc., can solve problems such as insufficient generalization ability, inability to adapt to retinal layer deformation and OCT image noise, low segmentation accuracy, etc. To achieve the effect of improving generalization ability and robust performance, improving generality, and improving accuracy

CN111583291AActive Publication Date: 2020-08-25SUN YAT SEN UNIV
7 Cites 9 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Publication Date
2020-08-25

Smart Images

  • Figure 1
    Figure 1
  • Figure 2
    Figure 2
  • Figure 3
    Figure 3
Patent Text Reader

Abstract

The invention discloses a layer segmentation method and system for a retina layer and an effusion area based on deep learning. The method comprises the following steps: acquiring a retina OCT data setof each node area in a medical system, dividing the data set into a pre-training data set and a test data set, and randomly translating data in the pre-training data set to obtain a training data set; carrying out forward propagation on the data in the training data set sent to the segmentation network in batches according to the constructed segmentation network and the corresponding loss function to obtain a segmentation prediction graph; according to a joint loss function formula, calculating a joint loss value between the segmentation prediction graph and a standard probability graph afterone-hot coding is carried out on the expert pixel-level marking image, carrying out back propagation on the joint loss value, and obtaining a segmentation network model through iterative training ofa preset period length; and testing the segmentation network model through the test data set to verify the reliability of the segmentation network model. According to the method and system, the generalization ability and the category segmentation accuracy of the segmentation network can be improved.
Need to check novelty before this filing date? Find Prior Art

Description

technical field

[0001] The present invention relates to the technical field of fundus image segmentation, in particular to a method and system for layer segmentation of retinal layers and effusion regions based on deep learning. Background technique

[0002] One of the most important structures in the eye, the retina is a very delicate and fragile tissue. Among many ophthalmic diseases, retinal diseases have always been the focus of research due to their high morbidity and blindness. Optical coherence tomography (OCT) is an imaging method of biological tissue, which has the characteristics of non-contact and non-invasive, high imaging speed and high resolution, and has been widely used to image retinal cross-sections.

[0003] At present, many diseases can cause complications of retinal diseases. For example, diabetes can cause diabetic macular edema (DME), that is, high blood sugar in diabetic patients can damage the retinal vascular epithelium and retinal fluid cells, cau...

Examples

Embodiment Construction

[0046] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments It is a part of embodiments of the present invention, but not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by persons of ordinary skill in the art without making creative efforts belong to the protection scope of the present invention.

[0047] see figure 1 , a method for layer segmentation of retinal layers and effusion regions based on deep learning proposed in the first embodiment of the present invention, which includes steps S101 to S104:

[0048] Step S101, obtain the retinal OCT data set of each node area in the medical system, divide the retinal OCT data set into a p...