R K W Schulze1, B d'Hoedt. 1. Department of Oral Surgery, Dental School, Johannes Gutenberg University, Augustusplatz 2, 55131 Mainz, Germany. rschulze@mail.uni-mainz.de
Abstract
OBJECTIVE: To develop an image analysis method for calculation of angular disparities between an object, temporarily equipped with a reference system, and a radiographic receptor. MATERIALS AND METHODS: A mathematical method based on a reference system containing two metallic spheres is developed, allowing calculation of inclination between the inter-spherical axis and the digital image receptor using image features. Experimental evaluation was done in standardized projection geometry for two sphere sizes at four randomly chosen inclinations per size, with each radiograph assessed three times. Truth was assessed threefold from photographs obtained at each inclination. RESULTS: Mean standard deviation between single assessments was 2.6 degrees. Significant differences (P Maloney/Rastogy=0.00) were found between absolute values of truth and calculated values (mean: -0.9 degrees; range: -6.0 degrees; 3.6 degrees), indicating a significant lack of accuracy. CONCLUSIONS: Although so far not sufficiently accurate, the method yields information relevant for correction of distortion in intra-oral radiology.
OBJECTIVE: To develop an image analysis method for calculation of angular disparities between an object, temporarily equipped with a reference system, and a radiographic receptor. MATERIALS AND METHODS: A mathematical method based on a reference system containing two metallic spheres is developed, allowing calculation of inclination between the inter-spherical axis and the digital image receptor using image features. Experimental evaluation was done in standardized projection geometry for two sphere sizes at four randomly chosen inclinations per size, with each radiograph assessed three times. Truth was assessed threefold from photographs obtained at each inclination. RESULTS: Mean standard deviation between single assessments was 2.6 degrees. Significant differences (P Maloney/Rastogy=0.00) were found between absolute values of truth and calculated values (mean: -0.9 degrees; range: -6.0 degrees; 3.6 degrees), indicating a significant lack of accuracy. CONCLUSIONS: Although so far not sufficiently accurate, the method yields information relevant for correction of distortion in intra-oral radiology.