Landmarking of Cephalograms


    Orthodontists have defined a certain number of characteristic points, or  landmarks, on the x-ray images of the human skull which are used to study  growth or as an diagnostic aid.  This work presents the first step towards the automatic extraction of these points.  They are defined with respect to  particular lines, which are first retrieved.  To begin with, the image is  preprocessed with a prefiltering operator (median filter) followed by edge detection (Mero-Vassy operator).  A knowledge-based line-following  algorithm is subsequently applied involving a production system with  organized sets of rules and a simple interpreter.  The a priori knowledge  implemented in the algorithm must take into account the fact that the lines  represent biological shapes; they can vary considerably from one patient to  the next.  The performances of the algorithm are judged with the help of  objective quality criteria of the image.  Having the exact shapes of the lines  then allows the computation of the positions of the landmarks.  This  processing was applied to seven x-rays as well as to artificially created  data.  The results depend greatly on the quality of the data.


The landmarks that the system has capacity to track are shown above.


  1. Levy-Mandel, A., Venetsanopoulos, A., Tsotsos, J., "A Methodology for Knowledge-Based Landmarking of Cephalograms", Computers and BioMedical Research 19, p282 - 309, 1986.

  2. Levy-Mandel, A., Tsotsos, J.K., Venetsanopoulos, A.N., "Knowledge-Based Landmarking of Cephalograms", Conference on Computer Assisted Radiology, Berlin, June 1985.

  3. Levy-Mandel, A., Venetsanopoulos, A., Tsotsos, J., "Knowledge-Based Image Processing with Application to the Landmarking of Cephalograms", Dixieme Collogue sur le Traitement du Signal et ses Applications, Nice, May 1985.

  4. Levy-Mandel, A.D., Tsotsos, J.K., Venetsanopoulos, A.N., "Knowledge Based Landmarking of Cephalograms", Proceedings of the 1984 Symposium on Communications, June 1984, Kingston, Ontario.

  5. Levy-Mandel, A.D., KNOWLEDGE-BASED LANDMARKING OF CEPHALOGRAMS, M.A.Sc. thesis,  Dept. of Electrical Engineering, University of Toronto, October 1984.