Translator Disclaimer
6 June 2000 Automatic evaluation of breast density for full-field digital mammography
Author Affiliations +
Women who have large breast area that is mammographically dense are greater risk of breast cancer than women with less mammographically dense breasts. It is a labor-intensive task to generate such a breast cancer risk information with the conventional screen film mammography. We have developed an automatic method to segment dense breast tissue areas using images acquired from full-field digital mammography systems. To evaluate its performance, a study was conducted to compare the segmentation results between the automatic method and a manually contouring method. A quantitative measurement, (delta) , is defined as the proportion of the segmented areas within a breast. The evaluation results indicate that there are 12 images whose (delta) difference between the automatic method and the manual method is less than 5%. Forty-nine images are between 5% and 10% and twenty images are within 15% in difference. On average, the process time required for the automatic method is approximately 18 seconds per image and 33 seconds per image for the manual method. The performance of our automatic method is comparable with the manual method. Yet, the automatic method does not require human intervention with the computer. We believe the automatic dense breast tissue segmentation method can be an effective tool to conduct studies of risk for breast cancer using FFDM images.
© (2000) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Shyhliang A. Lou and Yu Fan "Automatic evaluation of breast density for full-field digital mammography", Proc. SPIE 3979, Medical Imaging 2000: Image Processing, (6 June 2000);

Back to Top