Bone age assessment Application

    A fuzzy methodology is used to translate the natural language descriptions of the TW3 method for bone age assessment into an automatic classifier. The classifier is built upon a modified version of a fuzzy ID3 decision tree. No large data records are needed to train the classifier, i.e., to find out the classification rules, since the classifier is built upon rules given by the TW3 method. Only small data records are needed to fine-tune the fuzzy sets used to implement the rulebase.

    The ideas behind the procedure can be found in the following paper:

    Santiago Aja-Fernández; Rodrigo de Luis-García; Miguel Angel Martín-Fernández; Carlos Alberola-López, "A computational TW3 classifier for skeletal maturity assessment. A Computing with Words approach", Journal of Biomedical Informatics, vol. 37, pp. 99-107, 2004.[pdf]
    A MATLAB interface that implements the algorithm:
    Function tw3.zip
    User's manual: HOW_TO_USE.pdf