An Efficient Feature Extraction and Dyslexia Detection from Eye Movements Using Adaptive Neuro-Fuzzy Inference System (ANFIS)

Abstract

The major objective of the work is concern about the education community in their mission to be with the majority. The initial stage of the work is to study different mechanism for diagnosing learning disability (LD). Medical diagnosing searching becomes the major problem for movements of relationship among eye and dyslexia. To manage these problems several work have been demonstrated. But still it becomes most important difficulty how to extraction of best features from original data and how to improves the detection result in learning disability (LD). In order overcome these difficulties proposed an efficient wavelet transform (WT) for feature extraction and Adaptive Neuro-Fuzzy Inference System (ANFIS) based classifier dyslexia detection from eye movement signal. Eye movement of children was evaluated using Videooculographic (VOG) techniques. During this process both reading and non-reading tasks were measured based on eye movement of each children. After the measurement of eye movement then extract features using WT, then further the diagnosis of learning disability for classification using ANFIS system. Finally evaluate the performance of ANFIS classifier for dyslexia detection from eye movement signal. Compare to ANN based classification ,proposed ANFIS Classifier are less computational complication and detection result of methods are compared using parameters like accuracy, sensitivity and specificity, it shows that ANFIS Classifier are higher detection result than existing ANN classification