Detecting circumscribed lesions with the Hough transform

Conference · Thu Jan 11 04:00:00 EST 1996

OSTI ID:207503

Groshong, B R; Kegelmeyer, Jr, W P

We have designed and implemented a circumscribed lesion detection algorithm, based on the Hough Transform, which will detect zero or more approximately circular structures in a mammogram over a range of radii from a few pixels to nearly the size of the breast. We address the geometrical behavior of peaks in Hough parameter space (x,y,r) for both the true radius of a circular structure in the image (r = r{sub o}), and for the parameter r as it passes through this radius. In addition, we evaluate peaks in Hough parameter space by re-analyzing the underlying mammogram in the vicinity of the circular disk indicated by the peak. Discs suggested by the resulting peaks are accumulated in a feature image, scaled by a measure of their quality. These results are then rectified with respect to image contrast extremes and average value. The result is a feature with a continuously scaled pixel level output which suggests the likelihood that a pixel is located inside a circular structure, irrespective of the radius of the structure and overall mammogram contrast. These features are evaluated fast qualitative and quantitative performance metrics which permit circumscribed lesion detection features to be initially evaluated without a full end-to-end classification experiment.

• Research Organization: Sandia National Labs., Livermore, CA (United States)
• Sponsoring Organization: USDOE, Washington, DC (United States)
• DOE Contract Number: AC04-94AL85000
• OSTI ID: 207503
• Report Number(s): SAND--96-8498C; CONF-960219--1; ON: DE96006766
• Country of Publication: United States
• Language: English

Related Subjects

55 BIOLOGY AND MEDICINE
BASIC STUDIES
99 GENERAL AND MISCELLANEOUS
ALGORITHMS
BIOMEDICAL RADIOGRAPHY
IMAGE PROCESSING
MAMMARY GLANDS
PATHOLOGICAL CHANGES
TRANSFORMATIONS