Cone-beam reconstruction issues

From Peter's feedback, there are a number of issues with cone-beam:

• The scaling of the data, the result seems to be multiplied by the pixel_size_in_cm
• The full slice has to be reconstructed (end_x and end_y don't work)
• The slice is flipped up/down compared to Nabu; maybe ID16A negative rotation angles are not taken into account? The rotation direction really matters in cone beam geometry. Setting source_sample_dist to a negative works but is not elegant.
• horizontal translation movements are not taken into account
• FDK filter not applied anymore ? Discrepancy between 2024_07_24 and 2024_07_29 ?