Mammogram screening has been found an effective way for early detection and diagnosis, which can decrease breast cancer mortality significantly. Mass segmentation provides morphological features, which has important role for diagnosis. Here an end-to-end network for mammographic mass segmentation is proposed. The network employs a fully convolutional network (FCN) to model a potential function, followed by a conditional random fields (CRF) to perform structured learning. The network is designed to robustly Learn from a small dataset with poor contrast mammographic images. The two datasets used for model are from INBreast dataset and DDSM-BCRP dataset. Due to low contrast of mammographic images, image enhancement technique is used on the extracted ROI images followed by pixel position dependent normalization.
Input variables : Mammogram Image
Output Variables : Segmented Image
Visit Model : github.com
Additional links : arxiv.org
| Model Category | : | Public |
| Date Published | : | March, 2017 |
| Healthcare Domain | : |
Medical Technology
Provider |
| Code | : | github.com |
| Health Risk Management |
| Disease Detection |