Graphics Reference
In-Depth Information
specific parts of the visual word distribution which leads to the more discriminant
word-level representation. The precision-recall curves of these methods are presented
in Fig. 5. It indicates that our method overall outperforms the others.
5
Conclusion
In this paper, we propose a novel pathology image retrieval method for breast cancer.
Block LBP descriptor is used to describe the spatial characteristics of texture structure.
Then they are applied to generate into visual word representation by BoW scheme.
After low-rank and sparse composition operating, the word-level representation of each
image is decomposed into correlated part and specific part. Based on these two parts,
two pLSA models are leant to mine the high-level semantics existed in the images.
Finally, each image is represented by the combined topics of the two pLSA models.
Experiments on the pathology image database for breast cancer demonstrate the effec-
tiveness of our method. Further research will aim to apply Local Sensitive Hashing
(LSH) to boost the efficiency of retrieval when facing large database.
Acknowledgement. This work was supported by the National Natural Science Founda-
tion of China (No. 61371134), and the 973 Program of China (Project No.
2010CB327900).
References
1. Rebecca, S., Deepa, N., Ahmedin, J.: Cancer Statistics. CA Cancer Journal for Clini-
cians 63(1), 11-30 (2013)
2. Li, N., Zheng, R.S., Zhang, S.W., Zou, X.N., Zeng, H.M., Dai, Z., Chen, W.Q.: Analysis
and Prediction of Breast Cancer Incidence Trend in China. Chinese Journal of Preventive
Medicine 46(8), 703-707 (2012)
3. Fu, X.L.: The Atlas for Pathologic Diagnosis of Breast Tumours. Scientifics and Technical
Documents Publishing House, Beijing (2013)
4. Xue, Z.Y., Long, L.R., Antani, S., Thoma, G.R.: Pathological-based Vertebral Image Re-
trieval. In: IEEE International Symposium on Biomedical Imaging: From Nano to Macro
(ISBI), Chicago, pp. 1893-1896 (2011)
5. Lijia, Z., Shaomin, Z., Dazhe, Z., Hong, Z., Shukuan, L.: Medical Image Retrieval Using
Sift Feature. In: IEEE 2nd International Congress on Image and Signal Processing,
pp. 1-4. Tianjin (2009)
6. Caicedo, J.C., Izquierdo, E.: Combining Low-level Features for Improved Classification
and Retrieval of Histology Images. Transactions on Mass-Data Analysis of Images and
Signals 2(1), 68-82 (2010)
7. Marek, K., Paweł, F., Andrzej, O., Józef, K., Roman, M.: Computer-aided Diagnosis of
Breast Cancer Based on Fine Beedle Biopsy Microscopic Images. Computers in Biology
and Medicine 43(10), 1563-1572 (2013)
8. Ojala, T., Pietikäinen, M., Harwood, D.: A Comparative Study of Texture Measures with
Classification Based on Featured Distributions. Pattern Recognition 29(1), 51-59 (1996)
Search WWH ::




Custom Search