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.
You must be logged in to reply to this topic.