The method involves converting the picture into a digital signal. The signal is processed. It contains typical characteristics to identify the person. Many parts of the signal contain typical characteristics of the person. The signal amplitudes in different, non-neighbouring signal areas are compared. The filtering process involves using signal areas apart from each other. They are ordered in the form of parameters using a learning process. A minimum value of a function, involving the pattern signals and the signal group filtered is calculated. It is repeated for other signal groups during the learning process with the picture of the individual. The results stored in the memory provide recognition patterns.