Cephalogram analysis applying template matching and fuzzy logic

摘要

Cephalogram analysis has an important role in diagnosis, treatment and surgery of the skull and jaw. This field of science is mostly used by orthodontists. High speed and accuracy in automatic detection and geometrical evaluation and description of landmarks is widely demanded. An iterative self-organized data analysis technique using a combination of statistical pattern recognition and fuzzy logic-based decision making has been proposed as an efficient technique for detection of craniofacial landmarks. A large number of cephalograms has been considered in order to testify the performance of the system against different quality images. The images are primarily processed in order to enhance the regions of soft and hard tissues separately. Then for each landmark, different template matching plus fuzzy logic approach has been followed. Final landmarks are selected from a number of detected points by applying another fuzzy decision making criterion. The procedure is superior to the previous techniques for which either low accuracy has been reported or high quality images have been taken into consideration.