Abstract: Autism is a developmental disorder characterized among other early by alterations of socialization associated with a deficit of visual perception and/or aural and emotional expressions. To better understand the processes involved in autism, neurophysiologists analyzed responses to stimuli of autistic audio and video. The tools are commonly used fMRI, EEG and more recently "eye tracking". This device is simple to implement and use has begun to yield interesting results on the processes possibly involved in the perception of lack of photographs or films involving human presence. The study involving tools from Euclidean geometry and non-Euclidean, the trajectory of eye patients also showed interesting results. In this research, we want to confirm the results of the preliminary study but also going forward in understanding the processes involved in these experiments. Two tracks are followed, the first will concern the development of classifiers based on statistical data already provided by the system "eye tracking" and the second will be more focused on finding new descriptors from the eye trajectories. Regarding the classification several types of classification will be studied and implemented. The first classification study (the easiest) is to classify into two groups (people with autism and people without autism) results from the experiments. However, the test population is composed of more or less rehabilitated with autism, several groups will be proposed. The classifiers of the type "k-means", "neural networks", "SVM" will be tested in priority while knowing that other classifiers can be studied. The extraction parameters are most informative when studied in order to connect them with the processes involved in autism spectrum disorders. Concerning the second aspect of this research, it will be directed towards the search for new parameters from the analysis of the trajectory eye. Given the complex dynamics underlying the time series or trajectories, it is natural to turn to tools from the information theory or chaos theory. This assumption is realistic if we consider that the trajectory corresponds to the output of a nonlinear dynamic system excited by an input: the visual stimulus.