Improve image quality in computed utilizing redundant information in projection data sets

The invention relates to a method for detecting a movement (B) of a body (9) between the receiving points in time (t1, t2) of at least two recorded data sets (11), in a respective pair of comparison (21) of a first (i) and a second (j) recording data sets (11) a) for virtual cutting planes (22, 23) of the body (9) on the basis of the first receiving data set (i) a first intermediate function value (SNI) of attenuation values ​​of all (in the plane of section 22, located 23) body members (17) and based on the second recording record (j) a second intermediate function value (SNj) is determined of the attenuation values, and b) for each cross-sectional plane (22, 23) a difference value from the intermediate function values ​​(Sni, Snj) is determined and c) (for the two recorded data sets i, j) is a total error value (e The invention provides that the virtual cutting planes (22, 23) have a common line of intersection, and for the respective recording time point (t1, t2) the cutting lines (22 ', 23') pairs differed in at least one pixel (20).