Analysis of human eye movements during the plot inspection as a tool of assessment of local informative value of the 12-lead ECG

The main goal of every method of automatic analysis, recognition and also understanding of any medical signal is to follow the physician’s method of observation and thinking. Apparently it is easy, because the procedures of human interpretation of biosignals are now well standardized. In fact the real way of making observations and thinking about diagnosis highly depends on the observer experience and personal skills. In this paper, a new eyetrackbased approach is proposed for quantitative assessment of these important factors. The visual experiment carried out on cardiologists of various professional experience supplied the scanpaths data for the analysis in context of observed ECG traces The proposed approach, without the verbalization necessity, allows the cardiologists know-how to be extracted, analyzed and implemented in the automatic interpretation algorithm for better emulate the human way of thinking. In can be very useful for designing of new algorithms for automatic ECG signal analysis, recognition and also understanding by the machines.