Title :
Coding theoretic approach to image segmentation
Author :
Ndili, Unoma ; Nowak, Robert D. ; Figueiredo, Mario A.T.
Author_Institution :
Dept. of Electr. & Comput. Eng., Rice Univ., Houston, TX, USA
fDate :
6/23/1905 12:00:00 AM
Abstract :
This paper introduces multi-scale tree-based approaches to image segmentation, using Rissanen\´s coding theoretic minimum description length (MDL) principle to penalize overly complex segmentations. Images are modelled as Gaussian random fields of independent pixels, with piecewise constant mean and variance. This model captures variations in both intensity (mean value) and texture (variance). Segmentation thus amounts to detecting changes in the mean and/or variance. One algorithm is based on an adaptive (greedy) rectangular recursive partitioning scheme. The second algorithm is an optimally pruned "wedgelet" decorated dyadic partitioning. We compare the two schemes with an alternative constant variance dyadic CART (classification and regression tree) scheme which accounts only for variations in mean, and demonstrate their performance on SAR images
Keywords :
Gaussian distribution; adaptive estimation; image coding; image resolution; image segmentation; image texture; piecewise constant techniques; radar imaging; recursive estimation; synthetic aperture radar; tree data structures; Gaussian random fields; Rissanen MDL principle; SAR images; adaptive rectangular partitioning; coding theory; greedy recursive partitioning; image segmentation; independent pixels; intensity; minimum description length; multi-scale segmentation; optimally pruned dyadic partitioning; performance; piecewise constant mean; texture; tree-based segmentation; variance; wedgelet decorated dyadic partitioning; Codes; Decoding; Image segmentation; Image sequence analysis; Military computing; Parameter estimation; Partitioning algorithms; Pixel; Probability distribution; Telecommunication computing;
Conference_Titel :
Image Processing, 2001. Proceedings. 2001 International Conference on
Conference_Location :
Thessaloniki
Print_ISBN :
0-7803-6725-1
DOI :
10.1109/ICIP.2001.958055
