A new detection method for cognitive impairments is presented utilizing an eye tracking signals in a text reading test. This research enhances published articles that extract combination of various features. It does so by processing entire eye-tracking records either in time or frequency whereas applying only basic signal pre-processing. Such signals were classified as a whole by Convolutional Neural Networks (CNN) that hierarchically extract substantial features scatter either in time or frequency and nonlinearly binds them using machine learning to minimize a detection error. In the experiments we used a 100 fold cross validation and a dataset containing signals of 185 subjects (88 subjects with low risk and 97 subjects with high risk of dyslexia). In a series of experiments it was found that magnitude spectrum based representation of time interpolated eye-tracking signals recorded the best results, i.e. an average accuracy of 96.6% was reached in comparison to 95.6% that is the best published result on the same database. These findings suggest that a holistic approach involving small but complex enough CNNs applied to properly pre-process and expressed signals provides even better results than a combination of meticulously selected well-known features.

No abstract

This paper describes a method for dysgraphia disorder detection based on the classification of handwritten text. In the experiment we have verified proposed approach based on the conventional signal theory. Input data consists of the handwritten text by dysgraphia diagnosed children. Techniques for early dysgraphia detection could be applied in the schools to detect children with a possible diagnosis of dysgraphia and early intervention could improve their lives. The main goal of research is to develop a tool based on a machine learning for schools to diagnose dyslexia and dysgraphia. An experiment was performed on the dataset of 120 children in the school age (63 normally developing and 57 dysgraphia diagnosed). The main advantage is the simple algorithm for preprocessing of the raw data. Then was designed simple 3-layers convolutional neural network for classification of data. On the test data, our model reached accuracy 79.7%.

No abstract

No abstract