The objective of this bachelor thesis is to classify dyslexia disorder by analyzing eye-tracking data captured during text-reading tasks. In particular, the student will focus on encoding eye-tracking data into different forms of 2D images to compactly represent various characteristics of the captured eye movements. The image representations will encode information, such as raw eye-movement data, saccades, or fixations, and will be then classified by a suitable machine-learning method. Specifically, the parts of image representations (i.e., sub-images) will be classified separately and the ensemble approach will be applied. The proposed approaches will be experimentally evaluated on the publicly available ETDD70 dataset of eye-tracking data. The textual part will contain a survey of existing approaches for dyslexia classification, a description of the proposed image representations, and an evaluation of the experimental results.