Mass segmentation of mammograms using Markov models associated with constrained clustering
- PDF / 3,402,727 Bytes
- 21 Pages / 595.224 x 790.955 pts Page_size
- 29 Downloads / 167 Views
ORIGINAL ARTICLE
Mass segmentation of mammograms using Markov models associated with constrained clustering ´ Cruz-Barbosa1 Raul
1 · Luis Enrique Sucar2 ´ ´ · Saiveth Hernandez-Hern andez
Received: 10 January 2020 / Accepted: 22 June 2020 © International Federation for Medical and Biological Engineering 2020
Abstract In this paper, we propose four variants of the Markov random field model by using constrained clustering for breast mass segmentation. These variants were tested with a set of images extracted from a public database. The obtained results have shown that the proposed variants, which allow to include additional information in the form of constraints to the clustering process, present better visual segmentation results than the original model, as well as a lower final energy which implies a better quality in the final segmentation. Specifically, the centroid initialization method used by our variants allows us to locate about 90% of the regions of interest that contain a mass, which subsequently with the pairwise constraints helped us recover a maximum of 93% of the masses. The segmentation results are also quantitatively evaluated using three supervised segmentation measures. These measures show that the mass segmentation quality of the proposed variants, considering the breast density level, is consistent with the corresponding segmentation annotated by specialized radiologists. Keywords Mass segmentation · Markov random field model · Constrained clustering
1 Introduction Breast cancer is a global health problem that causes more than 500,000 deaths in the female population every year [14, 21, 48]. Although the use of mammograms has proven to be an effective tool for early detection and diagnosis of this cancer [16, 26], two challenges remain for radiologists. The former is related to the findings, since masses are difficult to detect due to their variation in characteristics such as size, shape, and density. The latter refers to the radiologists’ expertise, since it depends on their level of training and experience for mass detection. Consequently, several studies report that radiologists are, on average, 75% successful at performing a mammographic analysis for cancer detection [26]. A computer-aided detection (CADe) system can be used to assist radiologists in the detection of anomalies in Ra´ul Cruz-Barbosa
[email protected] 1
Applied Artificial Intelligence Laboratory, Computer Science Institute, Universidad Tecnol´ogica de la Mixteca, Huajuapan, Oaxaca, M´exico
2
´ Instituto Nacional de Astrof´ısica, Optica y Electr´onica, Tonantzintla, Puebla, M´exico
mammograms and can improve their effectiveness up to 82% [16, 43]. A CADe system consists of several stages [11, 44], where the image segmentation one is crucial for further stages of this system. For example, mass detection in mammograms is a problem that continues to be investigated. A given taxonomy for mass segmentation methods is shown in [16], where it is found that those based on clustering are the most used [38]. A specific subc
Data Loading...
