The paper presents the process and results of tests of the pattern recognition algorithm. The algorithm has been developed for the sensor to track the activity of the eyes. The study was conducted on a group of ten people. The group was selected, so that it was possible to determine the inuence of gender, age, eye color, contact lenses and glasses on the result of the algorithm. The object of the study was also to optimize the pattern recognition algorithm in terms of CPU utilization. For this purpose the duration of the algorithm steps were measured as a function of the allocation of tasks to multiple processor cores.