|
مقاله
|
Abstract
|
|
|
Title:
|
Modeling Effect of Higher-Order Aberrations on Visual Function in Keratoconic Eyes Using Adaptive Neural Network Based Fuzzy Inference System (ANFIS)
|
Author(s):
|
Aref Abedjooy, Mehdi Hosseini Tehrani, Amir Houshang Beheshtnejad
|
Presentation Type:
|
Oral
|
Subject:
|
Cornea and Anterior Segment
|
Others:
|
|
Presenting Author:
|
|
Name:
|
Aref Abedjooy
|
Affiliation :(optional)
|
-
|
E mail:
|
arefabedjooy@gmail.com
|
Phone:
|
|
Mobile:
|
09111434981
|
|
|
Purpose:
|
To investigate relationship between corneal higher-order aberrations and visual function in keratoconic eyes.
|
Methods:
|
This study included 87 eyes of 47 patients with keratoconus (KC). logMAR BCVA and AULCSF were obtained using CSV-1000E. Data of corneal wavefront HOAs related to KC subjects were recorded by a Tracey iTrace aberrometer. An ensemble Adaptive Neural Network Based Fuzzy Inference System (ANFIS) was developed using commercial software. Inputs to the ANFIS model consisted of HOAs which significantly correlated with logMAR BCVA and AULCSF. The data set was randomized; two third of the randomized data were used for training and the remaining one third was applied for testing. To evaluate performance of the proposed model, RMSE and MAPE of the predicted and observed values were used.
|
Results:
|
Eighty-seven eyes of 47 patients with keratoconus who aged 16 to 51 years old (27.02_+6.63) were included. Statistically significant correlation was found between logMAR and 3rd-order RMS(S3), 4th-order RMS(S4), 5th-order RMS(S5), vertical coma and horizontal coma. In addition, there were significantly negative correlations between AULCSF and S5, horizontal coma. Moreover, the related test results (RMSE and MAPE) were 0.0138 and 0.0151, respectively.
|
Conclusion:
|
From analysis of KC subjects, the pattern of visual function variations could be detected from HOAs with an ANFIS model.
|
Attachment:
|
|
|