Consistent performance measurement of a system to detect masses in mammograms based on blind feature extraction

BioMedical Engineering OnLine

Antonio García‐Manso¹  Carlos J García‐Orellana¹  Horacio González‐Velasco¹  Ramón Gallardo‐Caballero¹  Miguel Macías Macías¹ 

[1] Pattern Classification and Image Analysis Group, University of Extremadura, Avenida de Elvas s/n, Badajoz, Extremadura (SPAIN

关键词: DDSM;    Mammography;    Breast cancer;    CAD;    SVM;    NN;    ICA;   

Others  :  797964
DOI  :  10.1186/1475-925X-12-2

 received in 2012-10-30, accepted in 2012-12-26,  发布年份 2013

【 摘 要 】

Background

Breast cancer continues to be a leading cause of cancer deaths among women, especially in Western countries. In the last two decades, many methods have been proposed to achieve a robust mammography‐based computer aided detection (CAD) system. A CAD system should provide high performance over time and in different clinical situations. I.e., the system should be adaptable to different clinical situations and should provide consistent performance.

Methods

We tested our system seeking a measure of the guarantee of its consistent performance. The method is based on blind feature extraction by independent component analysis (ICA) and classification by neural networks (NN) or SVM classifiers. The test mammograms were from the Digital Database for Screening Mammography (DDSM). This database was constructed collaboratively by four institutions over more than 10 years. We took advantage of this to train our system using the mammograms from each institution separately, and then testing it on the remaining mammograms. We performed another experiment to compare the results and thus obtain the measure sought. This experiment consists in to form the learning sets with all available prototypes regardless of the institution in which them were generated, obtaining in that way the overall results.

Results

The smallest variation from comparing the results of the testing set in each experiment (performed by training the system using the mammograms from one institution and testing with the remaining) with those of the overall result, considering the success rate for an intermediate decision maker threshold, was roughly 5%, and the largest variation was roughly 17%. But, if we considere the area under ROC curve, the smallest variation was close to 4%, and the largest variation was about a 6%.

Conclusions

Considering the heterogeneity in the datasets used to train and test our system in each case, we think that the variation of performance obtained when the results are compared with the overall results is acceptable in both cases, for NN and SVM classifiers. The present method is therefore very general in that it is able to adapt to different clinical situations and provide consistent performance.

【 授权许可】

   
2013 García‐Manso et al.; licensee BioMed Central Ltd.

【 图 表 】

Figure 1.

Figure 2.

Figure 3.

Figure 4.

Figure 5.

Figure 6.

Figure 7.

Figure 8.

Figure 9.

Figure 10.

Figure 11.

【 参考文献 】

• [1]Serio GV, Novello AC: The advisability of the adoption of a law that would expand the definition of mammography screening to include the review of x‐ray examinations by use of a computer aided detection device. 2003.
• [2]Heath M, Bowyer K, Kopans D, Moore R, Kegelmeyer P: The digital database for screening mammography. In proceedings of the 5th International Workshop on Digital Mammography. Toronto, Canada; 2000:212-218.
• [3]Kopans D: Breast imaging. Philadelphia, PA 19106 USA: Lippincott Williams & Wilkins; 2007.
• [4]Oliver A, Freixenet J, Martí J, Pérez E, Pont J, Denton ER, Zwiggelaar R: A review of automatic mass detection and segmentation in mammographic images. Medical Image Anal 2010, 14:87-110.
• [5]Horsch A, Hapfelmeier A, Elter M: Needs assessment for next generation computer‐aided mammography reference image databases and evaluation studies. International J Comput Assisted Radiology and Surgery 2011, 6(6):749-767.
• [6]Hyvärinen A, Hurri J, Hoyer PO: Natural Image Statistics. A Probablisctic Approach to Early Computational Vision. London, England: Springer‐Verlag London Limited; 2009.
• [7]Anjali P, Ajay S: A review on natural image denoising using independent component analysis (ICA) technique. Adv Comput Res 2010, 2:06-14.
• [8]Campanini R, Dongiovanni D, Iampieri E, Lanconelli N, Masotti M, Palermo G, Riccardi A, Roffilli M: A novel featureless approach to mass detection in digital mammograms based on support vector machines. Physisc Med Biol 2004, 49:961-975.
• [9]Angelini E, Campanini R, Iampieri E, Lanconelli N, Masotti M, Roffilli M: Testing the performances of different image representations for mass classification in digital mammograms. Int J Modern Phys C [Comput Phys Physical Comput] 2006, 17:113-131.
• [10]Hong BW, Brady M: A Topographic Representation for Mammogram Segmentation. Lecture Notes Comput Sci 2879, 2003:730-737.
• [11]Bradski G, Kaehler A: Learning OpenCV: Computer Vision with the OpenCV Library. 1005 Gravenstein Highway North, Sebastopol, CA 95472: O’Reilly Media; 2008.
• [12]Jolliffe IT: Principal Component Analysis (2nd ed.).. 175 Fifth Avenue, New York, NY 10010, USA: Springer‐Verlag New York, Inc.; 2002.
• [13]Hyvärinen A, Karhunen J, Oja E: Independent Component Analysis. 605 Third Avenue, New York, NY 10158‐0012, USA: John Wiley & Sons; 2001.
• [14]Ripley B: FastICA Algorithms to perform ICA and Projection Pursuit. 2009. http://cran.r-project.org/web/packages/fastICA/fastICA.pdf webcite
• [15]Gonzalez RC, Woods RE: Digital image processing. Upper Saddle River, NJ: Prentice‐Hall; 2008.
• [16]Bishop CM: Pattern Recognition and Machine Learning. Secaucus, NJ, USA: Springer‐Verlag New York, Inc.; 2006.
• [17]Vapnik VN: The Nature of Statistical Learning Theory. 175 Fifth Avenue, New York, NY 10010, USA: Springer‐Verlag New York, Inc.; 2000.
• [18]Riedmiller H, Braun H: A direct adaptive method for faster backpropagation learning. The RPROP algorithm. IEEE Int Conference Neural Networks 1993, 1:586-591.
• [19]Zell A, Mache N, Huebner R, Mamier G, Vogt M, Schmalzl M, Herrmann K: SNNS (Stuttgart Neural Network Simulator). Neural Network Simulation Environments 1994, 254:165-186.
• [20]Chang CC, Lin CJ: LIBSVM: A library for support vector machines. ACM Trans Intelligent Syst Technol 2011, 2(3):1-27.
• [21]Hyvarinen A: Fast and robust fixed‐point algorithms for independent component analysis. IEEE Trans Neural Networks 1999, 10(3):626-634.
• [22]Fawcett T: An introduction to ROC analysis. Pattern Recognition Lett 2006, 27:861-874.
• [23]Gallas BD, Chan HP, D’Orsi CJ, Dodd LE, Giger ML, Gur D, Krupinski EA: Evaluating Imaging and Computer‐aided Detection and Diagnosis Devices at the FDA. Academic Radiology 2012, 19(4):463-477.
• [24]Lausser L, Kestler HA: Robustness Analysis of Eleven Linear Classifiers in Extremely High‐Dimensional Feature Spaces. In ANNPR, Volume 5998 of Lecture Notes in Computer Science. Edited by Schwenker F, Gayar NE. Springer; 2010:72-83.
• [25]Bowyer KW: Digital Image Database with Gold Standard and Performance Metrics for Mammographic Image Analysis Research. 1999.