Computer-assisted bone age assessment based on features automatically extracted from a hand radiograph

Comput Med Imaging Graph. 1995 May-Jun;19(3):251-9. doi: 10.1016/0895-6111(95)00005-b.

Abstract

This paper presents a computer-aided classification algorithm to assist the radiologist in the bone age assessment of pediatric patients. The classification is based on features automatically extracted from two regions of Computed Radiography (CR) left hand wrist images: phalangeal region of interest (PROI) and carpal bone region of interest (CROI). Due to imprecise nature of the bone age assessment problem, a fuzzy classifier for both regions has been developed. After defining a membership function for each region, features are processed yielding a matrix which maps the set of features to a year of age within the predefined range. The grades of membership are described as membership function values in the interval [0, 1]. A classification rule based on a max-sum operator, processes the matrix assessing the bone age. Since both regions are analyzed independently, two bone age assessments are obtained. They reflect the phalangeal and carpal bones maturity individually. In pathological cases the discrepancy between both assessments may reach as much as 2 yr.

Publication types

  • Research Support, U.S. Gov't, P.H.S.

MeSH terms

  • Age Determination by Skeleton*
  • Algorithms
  • Carpal Bones / diagnostic imaging
  • Carpal Bones / growth & development
  • Child
  • Cluster Analysis
  • Epiphyses / diagnostic imaging
  • Epiphyses / growth & development
  • Fingers / diagnostic imaging
  • Fingers / growth & development
  • Fuzzy Logic
  • Hand / diagnostic imaging*
  • Humans
  • Pattern Recognition, Automated
  • Radiographic Image Interpretation, Computer-Assisted*
  • Tomography, X-Ray Computed
  • Wrist / diagnostic imaging*