A Survey on Segmentation Techniques of Mammogram Images
Mammogram images are important tools allowing visualization of various types of breast cancer. In fact, cancer detection refers to the extraction of region of interest ROI, which represents the tumor, in the mammogram image. In medical imaging field, Comp
- PDF / 165,776 Bytes
- 12 Pages / 439.37 x 666.142 pts Page_size
- 45 Downloads / 182 Views
Abstract Mammogram images are important tools allowing visualization of various types of breast cancer. In fact, cancer detection refers to the extraction of region of interest ROI, which represents the tumor, in the mammogram image. In medical imaging field, Computer Aided Diagnosis systems (CAD) are used to analyze this type of images. To extract region of interest from mammograms, image segmentation methods have been wildly applied. These methods consist of partitioning the image on meaningful regions or segments easy to analyze. There are various techniques and methods of segmentation of mammogram images in the literature. In this paper, we present a survey of different approaches of segmentation that we compared theoretically in terms of advantages and drawbacks, particulary for mammogram images. Keywords Mammogram images ⋅ Image segmentation ⋅ Region of interest ROI ⋅ Image analysis
I. Ait lbachir (✉) ⋅ R. Es-salhi ⋅ I. Daoudi ⋅ S. Tallal ⋅ H. Medromi Systems Architecture, ENSEM, Hassan II University, Casablanca, Morocco e-mail: [email protected] R. Es-salhi e-mail: [email protected] I. Daoudi e-mail: [email protected] S. Tallal e-mail: [email protected] H. Medromi e-mail: [email protected] © Springer Science+Business Media Singapore 2017 R. El-Azouzi et al. (eds.), Advances in Ubiquitous Networking 2, Lecture Notes in Electrical Engineering 397, DOI 10.1007/978-981-10-1627-1_43
545
546
I. Ait lbachir et al.
1 Introduction Breast cancer is the second leading cause of mortality among women in the entire world, exceeded only by lung cancer. According to American Cancer Society, about 1 in 8 (12 %) women in the US will develop invasive breast cancer during their lifetime, and the foundation of Lalla Salma against the cancer affirms that 36.12 % of women in Morocco develop breast cancer. Only early detection can reduce the rate of mortality and increase recovery. Currently, mammography is the dominant tool to visualize and detect breast cancer, using low energy X-rays. The analysis of mammogram images remain a challenge among researchers, but in the few last decade, plenty of methods and Computer Aided Diagnosis systems (CAD) have been proposed. The goal of CAD systems is to assist and help radiologists in their interpretations and decision. We present in Fig. 1 the flowchart of mammogram image processing in CAD systems. In the first step, the image is pre-processed by removing noise and superfluous data, and then applying enhancement algorithms. The objective of this step is to prepare mammogram images to the following process. In the second step, titled segmentation, we aim to extract regions of Interest (ROI), which represent tumors, thereafter classified on benign or malignant masses. Once the mammogram segmentation is done, the goal of the third step is to describe the extracted ROI using different features like grey level histogram, intensity, size, texture and shape. The fourth step consist in selecting required features used in the final step to classify tumors as benign or malignan
Data Loading...
