A method for optimizing the interpretation of analog image signals or sequences of image signals output by medical image recording devices. The correlation of consecutively recorded image signals is tested and, based on the test, the image signals are identified as depicting the same or different images. More specifically, if the correlation is not less than a particular threshold value, it is established that the image signals depict the same image. If the correlation is less than a particular threshold value, it is established that the image signals possibly depict different images. Further, the threshold value is dynamically adjusted if the correlation has changed.