A hyper-surface-based modeling and correction of bias field in Mr images

Abstract EN

Dealing with the different artifacts in medical images is necessary to perform several tasks, including segmentation.

We introduce in this paper a novel method for bias field correction in Magnetic Resonance Imaging (MRI).

Using the segmentation results obtained by a modified Expectation Maximization clustering, the bias field is fitted as a hyper-surface in a 4D hyper-space.

Then, it is corrected based on the fact that voxels belonging to the same tissue should have the same intensity in the whole image.

So, after a quick and coarse unsupervised voxel labeling by clustering by parts is performed, the bias field is computed for reliably labeled voxels.

For the less reliably labeled voxels, the bias field is interpolated using a hyper-surface, estimated by a 4D Lagrangian interpolation.

We evaluated the efficiency of the proposed method by comparing segmentation results with and without bias field correction.

Also, we used the coefficient of variation within the MRI volume.

Segmentation results and the coefficient.