A volume data coding method based on the region segmentation is proposed. The volume data is three dimensional, thus a visualization method is needed. To visualize the desired portions in the volume data, region segmentation is indispensable. Hence, the region-oriented coding is suitable for the volume data. The finite mixture model is used to estimate the probability density function of the feature vectors obtained from the volume data. The volume data is segmented using a posteriori probability calculated from this finite mixture model. The information of the segmented regions are represented as follows: (i) the contours of the regions are represented by the chain code in each slice, (ii) the colors of the regions are represented as the coefficients of the polynomial which is used to approximate those colors, and (iii) since the posteriori probabilities can be regarded as the opacities of the voxels, the opacities which are needed to visualize the segmented regions are represented as the estimated parameters of the finite mixture model. Experimental results for CT data are shown.
[1]
Jeremy MG Taylor,et al.
Robust Statistical Modeling Using the t Distribution
,
1989
.
[2]
Thomas Engelhardt,et al.
Coding of arbitrarily shaped image segments based on a generalized orthogonal transform
,
1989,
Signal Process. Image Commun..
[3]
D. Rubin,et al.
Maximum likelihood from incomplete data via the EM - algorithm plus discussions on the paper
,
1977
.
[4]
Marc Levoy,et al.
Display of surfaces from volume data
,
1988,
IEEE Computer Graphics and Applications.
[5]
M. Kunt,et al.
Second-generation image-coding techniques
,
1985,
Proceedings of the IEEE.
[6]
R. Redner,et al.
Mixture densities, maximum likelihood, and the EM algorithm
,
1984
.
[7]
Naoyuki Ichimura,et al.
Inexhaustive region segmentation by robust clustering
,
1995,
Proceedings., International Conference on Image Processing.