Identifying cervical cancer lesions using temporal texture analysis

Cervical cancer is an important worldwide disease due to the high rate of incidence in the population. Several tests have been developed for detection of this illness; colposcopy is one of the diagnostic tests employed in recognition of lesions. Colposcopy performs a visual examination of the cervix based on temporal reaction of the surface stained with acetic acid, in order to identify lesions in cervix. In this paper we present an approach for identification of lesions based on temporal texture analysis in order to detect important patterns of lesions during colposcopy exam. Texture metrics based on spatial information are used in order to analyze temporally the acetic acid response and deduce appropriate signatures. Neural networks are used for classification process and results are presented using Receiver Operating Characteristic (ROC) analysis for visualizing classifier performance. Experimental results show a good accuracy in classification of lesions using temporal texture analysis.

[1]  Nasser Kehtarnavaz,et al.  Classification of breast mass abnormalities using denseness and architectural distortion , 2002 .

[2]  A. Rosenfeld,et al.  Visual texture analysis , 1970 .

[3]  Tom Fawcett,et al.  ROC Graphs: Notes and Practical Considerations for Researchers , 2007 .

[4]  M.F. Parker,et al.  Hyperspectral diagnostic imaging of the cervix: initial observations , 1998, Proceedings Pacific Medical Technology Symposium-PACMEDTek. Transcending Time, Distance and Structural Barriers (Cat. No.98EX211).

[5]  B. Pogue,et al.  Image analysis for discrimination of cervical neoplasia. , 2000, Journal of biomedical optics.

[6]  Isabelle Claude,et al.  Integrated color and texture tools for colposcopic image segmentation , 2001, Proceedings 2001 International Conference on Image Processing (Cat. No.01CH37205).

[7]  Abhijit S. Pandya,et al.  Pattern Recognition with Neural Networks in C++ , 1995 .

[8]  Leopoldo Altamirano Robles,et al.  Identifying precursory cancer lesions using temporal texture analysis , 2005, The 2nd Canadian Conference on Computer and Robot Vision (CRV'05).

[9]  A. Östör,et al.  Colposcopy, Cervical Pathology: Textbook and Atlas , 1984 .

[10]  Kunio Miyazawa,et al.  Initial neural net construction for the detection of cervical intraepithelial neoplasia by fluorescence imaging. , 2002, American journal of obstetrics and gynecology.

[11]  Linda G. Shapiro,et al.  Computer Vision , 2001 .

[12]  Kagan Tumer,et al.  Ensembles of radial basis function networks for spectroscopic detection of cervical precancer , 1998, IEEE Transactions on Biomedical Engineering.

[13]  Isabelle Claude,et al.  Contour features for colposcopic image classification by artificial neural networks , 2002, Object recognition supported by user interaction for service robots.