An evaluation of image analysis techniques as applied to thermographs used in breast cancer detection

Due to the lack of a formal theory for designing image analysis systems it has been necessary to develop and implement a package of computer programs for evaluating image analysis techniques and designing image analysis systems. The package allows the image analysis system designer to experiment with different algorithms and techniques on his data. This package is described and a classification of image analysis techniques is given. Finally the package and a class of image analysis algorithms are applied to a particular problem: the computer assisted interpretation of breast thermograms for use in the early detection of breast cancer'. It is an important problem to due the observation that breast cancer increases surface temperature and that its early detection increases the five year survival rate of sufferers. Thermography is somewhat nonspecific but may be effectively used as a screening technique. An assessment is made of the algorithms applied to breast thermograms and proposals for future research are given