Browse Prior Art Database

Publication Date: 2013-Sep-20
Document File: 9 page(s) / 4M

Publishing Venue

The Prior Art Database


The invention provides a technique to reduce edge enhancement artifact by extrapolating an image with a signal from a neighbor slice before smoothing. The technique solves edge enhancement artifact problem acquired by a correction method with low pass filter. During extrapolation, high frequency signal or boundaries are removed and placed with average of the neighbor slices. The extrapolated image is expected to preserve the smoothness of a sensitive profile at the boundaries to the regions with low signal. After extrapolation, background and remaining holes are further filled with mean value of non-background image.

This text was extracted from a Microsoft Word document.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 26% of the total text.



The invention generally relates to a magnetic resonance imaging (MRI) system and more particularly to a technique that reduces edge enhancement artifact in MRI image.


Signal intensity correction is an inevitable post processing step for magnetic resonance imaging (MRI) with a surface coil. The corrections allow demonstrating a way to correct image intensity without affecting image contrast. In signal intensity correction technique, the image is corrected by dividing an original image by a radiofrequency field (B1) profile. The technique assumes that the radiofrequency field (B1) profile is of low spatial frequency. Main problem with the technique includes that smoothing algorithm assumes that B1 profile varies smoothly across the whole image space. In contrast, the image intensity falls off sharply outside a boundary of an anatomy. Therefore, assumption of smooth field profile of the image intensity is not valid near the boundary. Image intensity in-homogeneity is perceived as a smooth variation of intensity across the image.

There are numerous techniques to correct the non-uniformity. The techniques are classified into two main categories which include non-retrospective and retrospective. Non-retrospective technique requires prior information to perform the correction. Non-retrospective technique includes phantom based technique, simulation technique, extra body coil scan and special acquisition scheme. Phantom based technique or various simulations do not require considering complex electromagnetic properties of human body. The correction based on the prior information is not accurate. Image from the body coil requires an extra scan which is not homogeneous at high fields due to wave behavior.

Retrospective technique doesn’t require prior information and are automated. Retrospective intensity correction techniques are based on either the tissue classification of desired true image or smoothness of the B1 field profile as a natural assumption is that the acquired image is true image multiplied by the B1 field profile of the coil or system. True image estimation requires statistics and iteration to acquire an optimized intensity histogram for the corrected image. Alternatively, the B1 field profile is estimated based on the image. The image is corrected by dividing the original image by B1 profile which is of low spatial frequency. However, the technique provides exhibition edge enhancement at the image boundary which is between image support and the background/holes in the image. Origin of such edge enhancement includes smoothing algorithms assuming that the signal modulator varies smoothly across whole image space. However, the image intensity fall off sharply outside the boundary of the anatomy and the assumption of smooth field profile of the image intensity is not valid near the boundary.

In order to reduce the edge enhancement, it is propos...