Feature Based Reading Skill Analysis Using Electrooculogram Signals

D'Souza, Sandra and Sriraam, N (2016) Feature Based Reading Skill Analysis Using Electrooculogram Signals. In: Advanced Computing and Communication Technologies,Advances in Intelligent Systems and Computing 452. Springer Science+Business Media, Singapore 2016, pp. 233-244.

[img] PDF
70.pdf - Published Version
Restricted to Registered users only

Download (650kB) | Request a copy


Developing reading skills is an individualistic characteristic and it differs from person to person. Since the eye movements can be intimately correlated to the reading process, by critical observation of the movement of the eyes reading process can be analysed and studied. This research work conducts a pilot study to propose and investigate the importance of eye movements in reading skill analysis. We have considered Electrooculogram (EOG) recorded signals from a group of ten healthy volunteers, of which are five normal readers and five poor readers. Simulation results show a classification accuracy of 67.7 and 88 % using Yule-Walker’s and Burg’s estimation methods respectively for horizontal EOG and 74 and 81 % for vertical EOG. The Burg’s estimation method stands out better for classification of reading skills. The results indicate the suitability of proposed scheme for identifying the poor readers and hence provide required assistance to people with reading disabilities.

Item Type: Book Section
Uncontrolled Keywords: EOG � Reading � AR model parameters � Classification � Artificial neural network
Subjects: Engineering > MIT Manipal > Instrumentation and Control
Depositing User: MIT Library
Date Deposited: 02 Sep 2017 06:27
Last Modified: 02 Sep 2017 06:27
URI: http://eprints.manipal.edu/id/eprint/149609

Actions (login required)

View Item View Item