A new method for extracting region of interest in mammograms

Author

Wen Lu ; Ruhai Dou ; Guangyu Zhang

Author_Institution

Radiol. Dept., Taishan Med. Coll., Taishan, China

fYear

2013

fDate

19-20 Oct. 2013

Firstpage

228

Lastpage

230

Abstract

In mammography computer-aided diagnosis, the automatic extraction of interesting region is one of the most difficult problems. This paper presents a method based on two-dimensional principal component analysis (2DPCA) to extract the region of interest (ROI) automatically. First, preprocess the mammograms, then, extract mammography features by 2DPCA method and edge-detection algorithm. Finally, extract ROI by neural network classifier. 60 cases were analyzed and 100 images which from Shandong medical imaging research institute were used in this investigation. The results show that a better positive detection ratio is obtained with this method. This approach can obtain better extraction accuracy by integrating 2DPCA, edge-detection algorithm and neural networks.

Keywords

edge detection; feature extraction; image classification; mammography; medical image processing; neural nets; principal component analysis; 2D principal component analysis; 2DPCA; Shandong Medical Imaging Research Institute; automatic region of interest extraction; computer aided diagnosis; edge detection algorithm; mammogram preprocessing; mammogram region of interest extraction; mammography feature extraction; neural network classifier; Classification algorithms; Computers; Feature extraction; Image edge detection; Neural networks; Principal component analysis; Training; Feature Extraction; Mammography; Region of Interest;

fLanguage

English

Publisher

ieee

Conference_Titel

Medical Imaging Physics and Engineering (ICMIPE), 2013 IEEE International Conference on

Conference_Location

Shenyang

Print_ISBN

978-1-4799-6305-8

Type

conf

DOI

10.1109/ICMIPE.2013.6864540

Filename

6864540