The research on this page is brought to you by Taylor & Francis Knowledge Centers. This collection is automatically generated from our most recent books and journals on this topic.
Breasts—composed primarily of fat and connective tissue—vary in form, shape, and size person to person, and often side to side. The anatomical boundaries of the breast are the second or third rib superiorly and the inframammary fold, the crease below the breast; the sternum medially and the mid-axillary line laterally (Pandya & Moore, 2011). Two regions of breast tissue overlay thoracic wall muscles: (1) the cone-shaped body of the breast with the nipple at its apex lies above the pectoralis major muscle and (2) an extension, the axillary tail, extends from the breast body into the armpit.
U-Net Convolutional Neural Networks for breast IR imaging segmentation on frontal and lateral view
Oliveira (2012) proposed an automatic segmentation method for IR lateral view images. First, the images are converted to greyscale for further background removal. After that, the authors detect the body edges as well as the inframammary fold for lower breast region removal. This method was evaluated using 328 images from DMR-UFF database and achieved an 93% accuracy, 95% sensitivity, and 96% specificity.