本研究提出全新的乳房攝影微鈣化區域之電腦輔助診斷系統，此方法可提高系統的靈敏度(真陽性率)與醫師們診斷時的能見度。本研究共有兩階段為重建與辨識微鈣化區域。
在第一階段重建疑似微鈣化區域中，利用Renyi’s資訊理論在小波轉換的不同分析層中將正常背景組織中的疑似微鈣化像素分割出來。然後藉由形態學運算與多數決法則以重建疑似微鈣化區域。此時，系統的靈敏度與每張影像的偽陽性值分別為97.19 %與22.88個。
在第二階段辨識微鈣化區域中，透過偏心率、緊密度、形狀慣量與灰階統計值等描述元，作為貝氏分類器與倒傳遞類神經網路分類疑似微鈣化區域的依據，從疑似微鈣化區域中擷取出49個描述元以辨識出微鈣化區域，並透過主成份分析找出有識別性的描述元，作為分類器的輸入訊號。
最後利用26張待研討的區域影像作為本研究方法的實證，此26張ROI(Region of Interest)皆為成大醫院之實際病歷，尤其當主成份數目選為9時，倒傳遞類神經網路的靈敏度與每張影像的偽陽性值分別為97.12 %與0.08個，其靈敏度將高出貝氏分類器約19 %左右，其差異最為明顯。

An innovative approach for computer-aided diagnosis algorithm upon microcalcification in digital mammograms is presented. It is efficient to increase the sensitivity (true positive detection rate) of the diagnosis system and enhance the visibility for medical doctors. This algorithm mainly consists of two stages of image processing including image reconstruction and extraction of microcalcification regions.
In order to reconstruct the microcalcification regions under suspicion, the potential microcalcification pixels are separated from the normal background tissue by employing Renyi’s information segmentation at different decomposition levels of the wavelet transform. Then the microcalcification regions under suspicion are reconstructed by the morphological operation and the majority vote rule. At this stage, the sensitivity is 97.19 % for segmentation with 22.88 false positive per image.
In order to extract the microcalcification regions under suspicion, Bayes classifier and Back-Propagation neural network are individually applied to classify the microcalcification regions by using linear combination of descriptors consisting of eccentricity, compactness, Shape Inertia and gray-level statistics. Microcalcification regions are detected by using a set of 49 descriptors. The discriminatory efficacy of these descriptors is verified by the approach of principal component analysis.
The presented algorithm is applied to a database of 26 regions of interest (ROI). In particular, the 97.12 % sensitivity is achieved at the cost of 0.08 false positive per image as the 9 principal components are used by the Back-Propagation neural network. To sum up, the sensitivity of the Back-Propagation neural network is approximately 20 % better than that obtained by Bayes classifier.

[1]http://www.doh.gov.tw/statistic/index.htm
[2]許慧貞、袁維新, ”乳房影像診斷一百例”,合記圖書出版
社,2006
[3]繆紹剛, ”數位影像處理 活用--Matlab,” 全華科技圖書
股份有限公司, 2002
[4]羅華強, ”類神經網路 Matlab的應用,” 高立圖書股份有
限公司, 2005
[5]C. K. Chui, “An Introduction to Wavelets,” Academic
Press, 1992.
[6]T. C. Wang and N. B. Karayianmis, “Detection of
Microcalcification in Digital Mammograms Using
Wavelets,” IEEE Trans. Med. Image., Vol. 17, No. 4,
pp. 498-509 , Aug. 1998.
[7]E. Sakka, A. Prentza and D. Koutsouris,
“Classification Algorithms for Microcalcifications in
Mammograms (Review),” Oncology Reports, Vol. 15,
pp. 1049-2056, 2006.
[8]K. Thangavel, M. Karnan, R. Sivakumar and A. Kaja
Mohideen, “Automatic Detection of Microcalcification
in Mammograms - A Review,” ICGST, 2006.
[9]H. D. Cheng, and H. Xu, “A Novel Fuzzy Logic Approach
to Mammogram Contrast Enhancement,” Information
Sciences Application, Vol 148, No. 1-4, pp. 167-184,
Dec. 2002.
[10]A. P. Dhawan, G. Buellton, R. Gordon, “Enhancement of
mammographic features by optimal adaptive neighborhood
image processing”, IEEE Trans. Med. Imag. , Vol 5,
No. 1, pp. 8-15, March, 1986.
[11]W. M. Morrow, R. B. Para jape, R. M. Rangayyan,
J. E. L. Desautels, “Region-based contrast
enhancement of mammograms”, IEEE Trans. Med. Imag.,
Vol. 11, No. 3, pp. 392–406, 1992.
[12]L. Shen, R. M. Rangayyan and J. E. L. Desautels,
“Application of Shape Analysis to Mammographic
Calcifications,” IEEE Trans. Med. Imag., Vol. 13,
No. 2, pp. 263-274, Jun. 1994.
[13]J. K. Kim, J. M. Park, K. S. Song, H. W. Park,
“Adaptive mammographic image enhancement using first
derivative and local statistics”, IEEE Trans. Med.
Imag., Vol 16, No. 5, pp. 495-502, 1997.
[14]B. Zheng, W. Qian, L. P. Clarke, “Digital
mammography: mixed feature neural network with
spectral entropy decision for detection of
microcalcifications”, IEEE Trans. Med. Imag., Vol.15,
No.5, pp. 589-597, Oct. 1996.
[15]S. Yu and L. Guan, “A CAD System for the Automatic
Detection of Clustered Mictocalcifications in
Digitized Mammogram Films,” IEEE Trans. Med. Imag.,
Vol. 19, No. 2, pp. 115-126 , Feb. 2000.
[16]A. Hojjatoleslami, L. Sardo, and J. Kittler, ”An RBF
Based Classifier for the Detection of
Microcalcification in Mammograms with Outlier
Rejection Capability,” IEEE Trans. Med. Imag.,
Vol. 21, No. 12,pp. 1379-1384, Jun. 1997.
[17]I. E1-Naqa, Y. Yang, M. N. Wernick, N. P. Galatsanos,
R. M. Nishikawa, “A Support Vector Machine Approach
for Detection of Microcalcifications”, IEEE Trans.
Med. Imag., Vol. 3, No. 12, Dec. 2002.
[18]H. P. Chan, C. J. Vyborny, H. McaMahon, C. E. Metz,
K. Doi, and E. A. Sickles, “Digital mammography: ROC
studies of the effects of pixel size and unship-mask
filtering on the detection of subtle
microcalcifications”, Invest. Radiol., Vol.22, No.7,
pp. 581–589, 1987.
[19]M. Witrh, M. Fraschini, and J. Lyon, “Contrast
Enhancement of Microcalcifications in Mammograms Using
Morphological Enhancement and Non-flat Structuring
Elements,” IEEE Symposium on CBNS’04., pp. 134-139,
Jun. 2004.
[20]R. N. Strickland and H. I. Hahn, “Wavelet Transform
for Detection Microcalcifications in Mammogramms,”
IEEE Trans. Med. Imag., Vol. 15, No. 2, pp. 218-229,
Apr. 1996.
[21]T. C. Wang and N. B. Karayianmis, “Detection of
Microcalcification in Digital Mammograms Using
Wavelets,” IEEE Trans. Med. Image., Vol. 17, No. 4,
Aug. 1998.
[22]M. J. Lado, P. G. Tahoces, A. J. Méndez, and
J. J. Vidal, “A Wavelet-Based Algorithm for Detecting
Clustered Microcalcifications in Digital Mammograms,”
Phys. Med. Biol., Vol. 26, No. 7, pp. 1294-1305,
July, 1999.
[23]S. Z. Hamid, R. R. Farshid, and P. N. D Siamak,
“Comparison of multiwavelet, wavelet, Haralick, and
shape features for microcalcification classification
in mammograms,” ELSEVIER Patten Recognition Society.,
Vol. 37, No. 10, pp. 1973-1986, Apr. 2004.
[24]A. Hojjatoleslami, L. Sardo, and J. Kittler, ”An RBF
based classifier for the detection of
microcalcification in mammograms with outlier
rejection capability,” IEEE Trans. Med. Image.,
Vol. 3, pp. 1379-1384, Jun. 1997.
[25]G. Boccignone, A. Chinanese and A. Picaroello,
“Computer Aided Detection of Microcalcifications in
Digital Mammograms,” Elsevier Science Med. Biol. Eng.
Computer, pp. 267-286, Apr. 2000.
[26]A. P. Dhawan, Y. Chitre, C. Kaiser-Bonasso and
M. Moskowitz, “Analysis of Mammographic
Microcalcifications Using Gray-Level Image Structure
Features.” IEEE Trans. Med. Imaging., Vol. 15, No. 3,
pp. 246-259, Jun. 1996.