Automatic Classification of Stromal Tissue in Prostate Images Based on Texture Descriptors

• Domingos L. L. Oliveira UFABC
• Marcelo Z. Nascimento UFU
• Leandro A. Neves UNESP
• Moarcir F. Godoy FAMERP
• Yan A. S. Duarte UFABC
• Pedro F. F. Arruda FUNFARME
• Dalisio S. Neto FUNFARME

Abstract

This paper presents an automatic classification system of stromal tissue in the prostate images based on texture descriptors. Normal, hyperplasia and cancer tissue class were evaluated. Initially, the segmentation of stromal tissue and extraction of regions of interest (ROI’s) was performed. Then, wavelet decomposition was applied to the ROI’s. Next, the Haralick and SVD descriptors were extracted. Feature selection based on genetic algorithm was applied. Finally, classification was performed using Random Forest algorithm. The area under the ROC curve (AUC) was calculated for normal versus cancer, normal versus hyperplasia and hyperplasia versus cancer, were achieved 0.962, 0.836 to 0.886, respectively.

References

Arivazhagan, S. and Ganesan, L. (2003). Texture segmentation using wavelet transform. Pattern Recognition Letters, 24(16):3197–3203.

Breiman, L. (2001). Random forests. Machine learning, 45(1):5–32.

Diamond, J., Anderson, N. H., Bartels, P. H., Montironi, R., and Hamilton, P. W. (2004). The use of morphological characteristics and texture analysis in the identification of tissue composition in prostatic neoplasia. Human Pathology, 35(9):1121–1131.

Doyle, S., Madabhushi, A., Feldman, M., and Tomaszeweski, J. (2006). A boosting cascade for automated detection of prostate cancer from digitized histology. In Medical Image Computing and Computer-Assisted Intervention–MICCAI 2006, pages 504–511. Springer.

Epstein, J. and Netto, G. (2007). Biopsy interpretation of the prostate. Lippincott Williams & Wilkins.

Gurcan, M., Boucheron, L., Can, A., Madabhushi, A., Rajpoot, N., and Yener, B. (2009). Histopathological image analysis: A review. Biomedical Engineering, IEEE Reviews in, 2:147–171.

Hall, M. A. (1999). Correlation-based feature selection for machine learning. PhD thesis, The University of Waikato.

Haralick, R. M., Shanmugam, K., and Dinstein, I. H. (1973). Textural features for image classification. Systems, Man and Cybernetics, IEEE Transactions on, (6):610–621.

INCA (2011). Estimativa 2012: Incidência de Câncer no Brasil. Technical report, Instituto Nacional de Câncer.

Junqueira, L. and Carneiro, J. (2008). Junqueira’s basic histology: text & atlas. Guanabara Koogan, 11 edition.

Ko, B. C., Kim, S. H., and Nam, J.-Y. (2011). X-ray image classification using random forests with local wavelet-based cs-local binary patterns. Journal of Digital Imaging, 24(6):1141–1151.

Mallat, S. (1999). A wavelet tour of signal processing.

Nguyen, K., Sabata, B., and Jain, A. K. (2012). Prostate cancer grading: Gland segmentation and structural features. Pattern Recognition Letters, 33(7):951–961.

Oliveira, D., Nascimento, M., Neves, L., Godoy, M., and Arruda, PFF, N. D. (2012a). Algoritmo para segmentação de lumens glandulares em tecidos da próstata. In Anais do XXIII Congresso Brasileiro em Engenharia Biomédica XXIII CBEB.

Oliveira, D., Nascimento, M., Neves, L., Godoy, M., and Arruda, PFF, N. D. (2012b). Segmentation of cell nuclei regions in epithelium of prostate glands. Journal of Modelling and Simulation of Systems, (3):21.

Paolone, D. R. (2010). Benign prostatic hyperplasia. Clinics in geriatric medicine, 26(2).

Pedrini, H. and Schwartz, W. R. (2008). Análise de imagens digitais: princípios, algoritmos e aplicações. Thomson Learning.

Ramos, R. P., Nascimento, M. Z. d., and Pereira, D. C. (2012). Texture extraction: An evaluation of ridgelet, wavelet and co-occurrence based methods applied to mammograms. Expert Systems with Applications.

Simon, C. P., Blume, L., and Doering, C. I. (2004). Matemática para economistas. Bookman.

Witten, I. H. and Frank, E. (2005). Data Mining: Practical machine learning tools and techniques. Morgan Kaufmann.