Cephalometric landmark identification using fuzzy wavelet edge detector

Cephalometric analysis is essential in acquiring medical measurement for analysis prior to treatment of orthognatic surgery, dentofacial growth and malocclusion. Manual tracing based on the visual assessment of radiographic image is a task that requires specially trained clinicians. We have proposed an automatic landmark identification technique which eliminates operator´s subjectivity, thereby improving the accuracy and efficiency of landmark identification. It is based on finer segmentation using template matching followed by fuzzy wavelet edge detection which determines the true edge. Finally the landmarks are marked on the edge on the basis of their geometric definitions. The landmarks identified were verified by orthodontists and the errors were found to be within the tolerance limit.