Graphics Reference
In-Depth Information
Pathology Image Retrieval by Block LBP Based pLSA
Model with Low-Rank and Sparse Matrix Decomposition
Yushan Zheng, Zhiguo Jiang, Jun Shi, and Yibing Ma
Image Processing Center, School of Astronautics, Beihang University
Beijing Key Laboratory of Digital Media
Beijing, China
yszheng@sa.buaa.edu.cn jiangzg@buaa.edu.cn
chris.shi331@gmail.com hemp110@126.com
Abstract. Content-based image retrieval (CBIR) is widely used in Computer
Aided Diagnosis (CAD) systems which can aid pathologist to make reasonable
decision by querying the slides with diagnostic information from the digital pa-
thology slide database. In this paper, we propose a novel pathology image re-
trieval method for breast cancer. It firstly applies block Local Binary Pattern
(LBP) features to describe the spatial texture property of pathology image, and
then use them to construct the probabilistic latent semantic analysis (pLSA)
model which generally takes advantage of visual words to mine the topic-level
representation of image and thus reveals the high-level semantics. Different
from conventional pLSA model, we employ low-rank and sparse matrix com-
position for describing the correlated and specific characteristics of visual
words. Therefore, the more discriminative topic-level representation corres-
ponding to each pathology image can be obtained. Experimental results on the
digital pathology image database for breast cancer demonstrate the feasibility
and effectiveness of our method.
Keywords: Image retrieval, computer aided diagnosis, breast cancer,
probabilistic latent semantic analysis, low-rank and sparse matrix composition.
1
Introduction
Computer Aided Diagnosis (CAD) system for breast cancer has attracted more and
more attention due to morbidity increase of breast cancer in female [1, 2]. Although new
technologies for breast cancer diagnosis have developed rapidly in the past few years,
the final diagnosis still relies on the pathological theories [3]. As the digital pathology
slides spread, senior pathologists can mark the lesion area with detailed descriptions on
the digital slides and share it to others through CAD systems or the Internet. In the other
hand, junior pathological doctors can get valuable suggestions by searching slides that
contain diagnosis information when facing indeterminable cases. Therefore, CAD sys-
tems consisting of pathology slide database with confirmed diagnosis information are
urgently required. But it is challenging to retrieve useful slides from the enormous data-
base effectively and accurately for the reason that the resolution of digital pathology
 
Search WWH ::




Custom Search