A dynamical appearance model based on multiscale sparse representation: segmentation of the left ventricle from 4D echocardiography

Med Image Comput Comput Assist Interv. 2012;15(Pt 3):58-65. doi: 10.1007/978-3-642-33454-2_8.

Abstract

The spatio-temporal coherence in data plays an important role in echocardiographic segmentation. While learning offline dynamical priors from databases has received considerable attention, these priors may not be suitable for post-infarct patients and children with congenital heart disease. This paper presents a dynamical appearance model (DAM) driven by individual inherent data coherence. It employs multi-scale sparse representation of local appearance, learns online multiscale appearance dictionaries as the image sequence is segmented sequentially, and integrates a spectrum of complementary multiscale appearance information including intensity, multiscale local appearance, and dynamical shape predictions. It overcomes the limitations of database-driven statistical models and applies to a broader range of subjects. Results on 26 4D canine echocardiographic images acquired from both healthy and post-infarct subjects show that our method significantly improves segmentation accuracy and robustness compared to a conventional intensity model and our previous single-scale sparse representation method.

Publication types

  • Research Support, N.I.H., Extramural

MeSH terms

  • Cardiac-Gated Imaging Techniques / methods*
  • Computer Simulation
  • Echocardiography, Three-Dimensional / methods*
  • Humans
  • Image Enhancement / methods
  • Image Interpretation, Computer-Assisted / methods*
  • Models, Cardiovascular*
  • Reproducibility of Results
  • Sensitivity and Specificity
  • Ventricular Dysfunction, Left / diagnostic imaging*