Image Quality in Digital Mammography

The substitution of digital representations for standard film/screen mammography analog images provides access to methods for digital storage and transmission and enables the use of a variety of digital image processing techniques, including enhancement and computer assisted detection and classification of abnormalities. Lossy image compression can further improve the efficiency of transmission and storage and can facilitate subsequent image processing. Digitization, digital acquisition, lossy compression, and enhancement and highlighting schemes all alter an image from its traditional form and it is critically important that any such alteration be clearly demonstrated to improve or at least not damage the utility of the image. The primary goal of this project was the development of a clinical experimental protocol for evaluating the comparative quality of analog, digital, and lossy compressed digital mammograms in a manner not suffering from the fundamental flaws of traditional receiver operating characteristic (ROC) methods. Such a protocol was developed and implemented in a pilot experiment involving 59 patients and 6 radiologist judges interpreting original, 12 bits per pixel (bpp) digitized, and lossy compressed digital mammograms down to .15 bpp. The protocol, experiment, and results are described along with related algorithms for compression and combined compression and