Nonlinear Anisotropic Filtering of MRI Data

Despite significant improvements in image quality over the past several years, the full exploitation of magnetic resonance image (MRI) data is often limited by low signal-to-noise ratio (SNR) or contrast-to-noise ratio (CNR). In implementing new MR techniques, the criteria of acquisition speed and image quality are usually paramount. To decrease noise during the acquisition either time averaging over repeated measurements or enlarging voxel volume may be employed. However, these methods either substantially increase the overall acquisition time or scan a spatial volume in only coarse intervals. In contrast to acquisition-based noise reduction methods we propose a postprocess based on anisotropic diffusion. Extensions of this new technique support 3-D and multi-echo MRI, incorporating higher spatial and spectral dimensions. The procedure overcomes the major drawbacks of conventional filter methods, namely the blurring of object boundaries and the suppression of fine structural details. The simplicity of the filter algorithm enables an efficient implementation even on small workstations. We demonstrate the efficient noise reduction and sharpening of object boundaries by applying this image processing technique to 2-D and 3-D spin echo and gradient echo MR data. The potential advantages for MRI, diagnosis and computerized analysis are discussed in detail.