Related Experiment Videos
SPASM: a 3D-ASM for segmentation of sparse and arbitrarily oriented cardiac MRI data
Hans C van Assen1, Mikhail G Danilouchkine, Alejandro F Frangi
1Division of Image Processing, Department of Radiology, Leiden University Medical Center, Albinusdreef 2, PO Box 9600, 2300 RC, Leiden, The Netherlands. h.c.v.assen@tue.nl
Medical Image Analysis
|January 28, 2006
Summary
A new technique, Sparse PAtch-based Segmentation (SPASM), accurately segments cardiac MRI data even with limited views. This method offers reliable cardiac volume calculation with reduced image data requirements.
Area of Science:
- Medical Imaging
- Computational Biology
- Cardiovascular Imaging
Background:
- Accurate segmentation of cardiac magnetic resonance imaging (MRI) data is crucial for quantitative analysis.
- Existing methods often struggle with undersampled or multi-planar datasets.
Purpose of the Study:
- To introduce and evaluate SPASM (Sparse PAtch-based Segmentation), a novel 3D active shape model technique for cardiac MRI segmentation.
- To assess SPASM's performance with sparse and arbitrarily oriented image data.
Main Methods:
- SPASM utilizes a two-stage iterative process for updating landmark positions, propagating information from image intersections to sparse regions.
- Feature point detection employs a fuzzy inference system based on fuzzy C-means clustering.
- Grid computing was used for parameter optimization and computational load distribution.
Main Results:
- SPASM achieved (sub-)pixel accuracy across various data sparsities (2-11 slices) and orientations.
- While performance differences were noted for sparse short-axis data (<6 slices), volume differences were not clinically relevant (<4 ml).
- SPASM demonstrated comparable or superior performance to other 3D model-based methods using significantly less image data.
Conclusions:
- SPASM provides accurate and robust automatic segmentation of cardiac MRI data, even with substantial undersampling.
- The method is relatively insensitive to initial model placement, offering a practical advantage.
- SPASM represents an efficient advancement in cardiac image analysis, requiring less data than existing techniques.