Optical system based data classification for diabetes retinopathy detection using machine language with artificial intelligence

doi:10.1007/s11082-023-05193-x

Optical system based data classification for diabetes retinopathy detection using machine language with artificial intelligence

2023-08-02

1

PRE

AI

S

Suraj Malik *

S

Sudharsanan Srinivasan

C

Chandra Shekhar Rajora

DOI：10.1007/s11082-023-05193-x

Original

Original request for help

Save

Abstract

En 中文

Diabetes causes DR. Diabetes duration influences retinopathy development. The retinal vein weakening may have no side effects or a little eyesight impairment at initially. Blindness may occur. DR intervention and treatment need early clinical indications. Thus, frequent eye examinations must guide patients to a doctor for a full eye inspection and therapy to avoid irreversible vision loss. This work develops a machine learning-based optical image-based data classification method for diabetic retinopathy identification. OCT analyses the retinal picture and the ensemble pulse coupled filtering and green histogram channel equalization-based adaptive filtering segment this picture for blood vessel characterization. CenterResnet-50 classifies images for color fundus detection. Classification accuracy, sensitivity, specificity, AUC, and ROC curves were examined for various optical retina pictures. The proposed method has 98% classification accuracy, 67% sensitivity, 73% specificity, and 63% AUC.

Keywords:

Optical image

Data classification

Diabetes retinopathy

Machine learning techniques

Optical coherence

Optical system based data classification for diabetes retinopathy detection using machine language with artificial intelligence

Abstract

Journal

Organization