Improving mass discrimination in mammogram-CAD system using texture information and super-resolution reconstruction
- PDF / 1,172,738 Bytes
- 10 Pages / 595.276 x 790.866 pts Page_size
- 62 Downloads / 197 Views
ORIGINAL PAPER
Improving mass discrimination in mammogram‑CAD system using texture information and super‑resolution reconstruction Sawsen Boudraa1 · Ahlem Melouah1 · Hayet Farida Merouani1 Received: 20 October 2018 / Accepted: 14 December 2019 © Springer-Verlag GmbH Germany, part of Springer Nature 2020
Abstract Screening helps to reduce mortality in the breast cancers. Mammography is a screening procedure used to detect breast cancer at an early stage. Computer-aided detection (CAD) systems can help in mammograms examination. Automatic differentiation between benign and malignant mammographic masses is a challenging task, due to high variability in mass structures. That is why, CAD systems frequently misdiagnose breast cancer. This paper presents a new CAD approach for mass detection in digital mammograms. The purpose of the proposed approach is to improve the discrimination between benign mass and malignant mass by reinforcing their statistics texture features. To achieve this aim, a new step based on super-resolution reconstruction is added to multistage CAD system. The proposed approach gives very good results comparing to other recent works. It achieves 96.7% classification accuracy using the MIAS (Mammography Image Analysis Society) dataset. This work shows that a super-resolution based approach improves the performance of the evaluated texture methods and thus outperforms benign/malignant mass classification for digital mammograms. Keywords Classification · Mammograms · Pattern recognition · Super-resolution · Texture features
1 Introduction Breast mass is an abnormal growth of cells in breast tissues. Breast mass can be classified into two categories, benign (not cancerous) or malignant (is potentially cancerous) (Hedawoo et al. 2016). A cancerous mass called also a tumor is an uncontrollable change and spread of breast cells. If the tumor is small, breast cancer is easily treated but if the tumor is large, breast cancer can result in death. The prognosis of invasive breast cancer is strongly influenced by the stage of the disease. Electronic supplementary material The online version of this article (https://doi.org/10.1007/s12530-019-09322-4) contains supplementary material, which is available to authorized users. * Sawsen Boudraa [email protected] Ahlem Melouah [email protected] Hayet Farida Merouani [email protected] 1
Department of Computer Science, Laboratory of Research in Computer Science (LRI), University of Badji Mokhtar, P.O. Box 12, 23000 Annaba, Algeria
Early stage detection of breast cancer is important as it can substantially reduce mortality rates (Oeffinger et al. 2015). Screening examination offers the opportunity to detect breast cancers early, when treatment is more likely successful. There are several imaging techniques for the breast examination, such as magnetic resonance imaging, ultrasound imaging, X-ray imaging, etc. Mammography is a special X-ray procedure for detecting breast cancer. Mammography is the best and cheapest way, with fewer side effects i
Data Loading...
