No abstract

Rough 3D data images obtained by computed tomography or magnetic resonance imagery are inadequate: this paper proposes a highlevel data structure called ellipsoidal skeleton. It is based on a tree of best partitions of the points set and features data compression, multi-level representation capabilities, surface reconstruction, interactive visualization, relevant parameters extraction, automatic matching and recognition.

Bone maturation is the only reliable indicator of growth and its radiologic assessment with or without automated systems is a qualitative method. Image processing allows the study of bone maturation with quantitative data. Carpal bone maturation was studied in 20 children (13 boys and 7 girls, ages ranging from 4 to 15 years) without any clinical evidence of endocrine disease by image analysis from computed tomography (CT) scans. Each wrist CT scan was processed in order to extract the carpal bones and to measure quantitative data regarding volume, axes of inertia and density for each bone. The volumes and the length of the inertia axes were significantly correlated with age. Whatever the age, there were strong correlations between the volume or the length of the main inertia axis of one carpal bone and that of all others. The decrease in the carpal bone volume measured from the processing procedure compared with the theoretical volume of bone defined from the length of the three inertia axes indicated a change in bone shape during growth. Although the mean density was constant, there was an increase in the standard deviation of density with age. Skeletal maturity assessment with image analysis from CT scans seems to be a good complementary investigation to determine bone age in children.

No abstract