Name
Abstract | ||
|---|---|---|
We have developed a computer-aided diagnosis scheme for detection of clustered microcalcifications at the early stage in mammograms. In the proposed method, the mammogram image is input to a filter bank for detecting nodular/line patterns. Then, the nodular pattern-enhanced image, the nodular/line pattern-enhanced image, and the second-order difference images for horizontal, vertical, and diagonal directions are generated at each resolution level. Next, the mammogram image is partitioned into regions of interest (5 mm × 5 mm). For each region of interest, based on the images generated from the filter bank at each resolution level, the objective features are extracted and used to discriminate the region of interest containing the clustered microcalcifications and the region of interest for normal tissue. Finally, discriminant analysis using these objective features is employed for detecting the clustered microcalcifications. A detection experiment was performed using 59 mammograms containing 106 clustered microcalcifications at the early stage; the sensitivity was 94.3% (100/106 clustered microcalcifications), the number of false positives was 0.051 per image, indicating that the proposed method is useful in detecting clustered microcalcifications at the early stage. © 2005 Wiley Periodicals, Inc. Syst Comp Jpn, 36(5): 68–79, 2005; Published online in Wiley InterScience (). DOI 10.1002/scj.20172 |
Year | DOI | Venue |
|---|---|---|
2005 | 10.1002/scj.v36:5 | Systems and Computers in Japan |
Keywords | Field | DocType |
filter bank,hessian matrix | Diagonal,Computer vision,Computer science,Filter bank,Computer-aided diagnosis,Hessian matrix,Artificial intelligence,Region of interest,Linear discriminant analysis,False positive paradox | Journal |
Volume | Issue | Citations |
36 | 5 | 0 |
PageRank | References | Authors |
0.34 | 2 | 5 |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Ryohei Nakayama | 1 | 14 | 4.80 |
| Yoshikazu Uchiyama | 2 | 69 | 9.58 |
| Koji Yamamoto | 3 | 1 | 1.02 |
| Ryoji Watanabe | 4 | 0 | 1.01 |
| Kiyoshi Namba | 5 | 2 | 1.38 |