Original Jpn J Radiol Technol 2000 ;
Inclusive Integral Evaluation for Mammograms Using the Hierarchical Fuzzy Integral (HFI)Model

TAKASHI AMANO, KAZUYA YAMASHITA, SHINICHI ARAO,
AKIRA KITAYAMA, AKIKO HAYASHI, SHINJI SUEMORI,1)
and YASUHIKO OHKURA2)
Department of Radiological Technology, Kawasaki College of Allied Health Professions
1)Department of Radiology, Kawasaki Medical School Hospital
2)Graduate School of Medicine, Course of Health Sciences, Osaka University
Received Jan. 25, 1999; Revision accepted Oct. 25, 1999; Code No. 521

Summary
Physical factors (physically evaluated values)and psychological factors (fuzzy measurements)of breast x-ray images were comprehensively evaluated by applying breast x-ray images to an extended stratum-type fuzzy integrating model. In addition, x-ray images were evaluated collectively by integrating the quality (sharpness, graininess, and contrast)of x-ray images and three representative shadows (fibrosis, calcification, tumor)in the breast x-ray images. We selected the most appropriate system for radiography of the breast from three kinds of intensifying screens and film systems for evaluation by this method and investigated the relationship between the breast x-ray images and noise equivalent quantum number, which is called the overall physical evaluation method, and between the breast x-ray images and psychological evaluation by a visual system with a stratum-type fuzzy integrating model. We obtained a linear relationship between the breast x-ray image and noise-equivalent quantum number, and linearity between the breast x-ray image and psychological evaluation by the visual system. Therefore, the determination of fuzzy measurement, which is a scale for fuzzy evaluation of psychological factors of the observer, and physically evaluated values with a stratum-type fuzzy integrating model enabled us to make a comprehensive evaluation of x-ray images that included both psychological and physical aspects.

Key words: Fuzzy measure, Fuzzy integral, Hierarchical fuzzy integral, Inclusive integral evaluation, Image quality