A non-invasive method is described for identifying contours with complex structures having an error recognition rate tending towards zero, which method can be used to distinguish and identify features of a slaughtered animal body with respect to each other, said method being utilized on the basis of a characteristic piece in particular to determine, by means of image processing, data for calculating the muscle-meat percentage, the trade classification and the associated trade value and market value, and also to rate the quality of slaughtered animal bodies, taking into account legal requirements.In accordance with the invention, during error recognition of contours of a slaughtered animal body which have complex structures and which are to be distinguished and identified with respect to each other, an operator at a workstation in the form of a PC interacts with the computer and provides sufficient indications relating to the unidentified or incorrectly identified contour to be sought, in that by employing an image reproduction device at least one support point is in each case set in the illustration of an image of the region of interest of the slaughtered animal body in one or several regions, in which data is to be determined from measurements.