Academic Commons

Presentations (Communicative Events)

Evaluation of Contrast Enhancement by Digital Equalization in Digital Mammography

Huda, Walter; Jin, Yinpeng; Laine, Andrew F.

Purpose: This study evaluated an algorithm based on a method of contrast enhancement by digital equalization (CEDE). Method: The algorithm was designed to enhance image contrast by employing digital equalization of digital mammograms. The CEDE algorithm was tested using ten mammograms with cancer (13 lesions) taken the University of South Florida data base, together with eight mammograms which only contained benign lesions. Three readers compared the processed images with the original mammograms for lesion conspicuity. A five point ranking scale was employed where a score of 3 corresponded to equal lesion visibility, ranks > 3 corresponded to superior lesion visibility, whereas ranks < 3 corresponded to markedly inferior lesion visibility. Results: The mean observer score for all lesions was always at least equal to that of the original digital mammogram (i.e., 3 or greater), and there was no evidence of any image distortion or other image processing artefacts. The mean rank (± standard deviation) for the 13 malignant lesions was 3.52 ± 0.38. The corresponding rank for the eight benign lesions was 3.33 ± 0.26. These differences were statistically significant in terms of standard error. Conclusion: The CEDE algorithm is capable of significantly enhancing lesion contrast in digital mammograms and our preliminary results indicate that this algorithm merits additional refinement and further (objective) evaluation.

Files

More About This Work

Academic Units
Biomedical Engineering
Published Here
August 20, 2010

Notes

Presented at the World Congress on Medical Physics and Biomedical Engineering, Chicago, July 23-28, 2000.

Academic Commons provides global access to research and scholarship produced at Columbia University, Barnard College, Teachers College, Union Theological Seminary and Jewish Theological Seminary. Academic Commons is managed by the Columbia University Libraries.