Integrating Geometric Configuration and Appearance Information into a Unified Framework for Anatomical Landmark Localization

A large number of medical image analysis applications rely on automatic anatomical landmark localization algorithms as a preliminary step for, e.g. segmentation based on deformable and statistical shape models (Heimann and Meinzer, 2009; Zhang et  al., 2012; Lay et al., 2013), registration of images using rigid (Hajnal et al., 2001) and deformable transformations (Johnson and Christensen, 2002; Urschler et al., 2006), construction of anatomical atlases for population studies (Toews et al., 2010), or to focus on anatomical structures of interest for computer-aided diagnosis (Doi, 2007) and regression tasks like skeletal age estimation (Thodberg et al., 2009; Štern et al., 2014).
Source: Medical Image Analysis - Category: Radiology Authors: Source Type: research