| 摘要 |
PROBLEM TO BE SOLVED: To enable identification of wrong connection of electrodes of right and left arms with a higher sensitivity without impairing identifying performance by providing a control unit with an artificial neural network to determine whether an electrode incorrectly attached to a patient exists or not. SOLUTION: An electric signal generated from a heart is picked up by electrodes 6A-6J and supplied to a signal processing unit 14 of an ECG analyzer 4, where the signal arriving is re-divided into six channels made up of electrodes 6A, 6B and 6C. The processed signals are sent out to an analysis unit 16 to discriminate various signals parameters of measurement signals. A selected one of the parameters is transmitted to a control unit 18 to be used as an input signal for first, second and third neural networks 18A, 18B and 18C. The three neural networks 18A-18C are operated to learn how to identify various wrong connections of the electrodes 6A-6J. |
