Minimum Description Length Shape and Appearance Models
The Minimum Description Length (MDL) approach to shape modelling is reviewed. It solves the point correspondence problem of selecting points on shapes defined as curves so that the points correspond across a data set. An efficient numerical implementation is presented and made available as open source Matlab code. The problems with the early MDL approaches are discussed. Finally the MDL approach is extended to an MDL Appearance Model, which is proposed as a means to perform unsupervised image segmentation.
KeywordsShape Modelling Minimum Description Length Appearance Models Point Correspondence Problem Unsupervised Vision Image Segmentation
Unable to display preview. Download preview PDF.
- 2.Cootes, T.F., Edwards, G.J., Taylor, C.J.: Active appearance models. In: Proc. Euro-pean Conf. on Computer Vision, vol. 2, pp. 484–498. Springer, Heidelberg (1998)Google Scholar
- 7.Minka, T.P.: Automatic choice of dimensionality for PCA, Technical Report 514, MIT Media Laboratory (2000)Google Scholar
- 8.Thodberg, H.H.: Hands-on Experience with Active Appearance Models, Medical Imaging 2002: Image Proc. In: Sonka, Fitzpatrick (eds.) Proc. SPIE, vol. 4684, pp. 495–506 (2002)Google Scholar
- 9.Cootes, T.F., Taylor, C.J.: On representing edge structure for model matching. In: Proc. IEEE CVPR, vol. 1, pp. 1114–1119 (2001)Google Scholar
- 10.Stegmann, M.B.: Analysis and Segmentation of Face Images using Point Annotations and Linear Subspace Techniques, IMM Technical Report IMM-REP-2002-22 (2002)Google Scholar
- 11.Davies, R.H.: thesis, Univ. of Manchester (2002), http://www.cfn.unimelb.edu.au/rhhd/