Magnetic resonance imaging using pattern recognition

Abstract

An MRI technique in which the similarity of samples from different portions of a body is determined and displayed. In one embodiment, the method can be used to track the spread of a known primary tumor to other portions of a patient's body. The MRI apparatus is used to produce a training set comprising one or more training samples. The training set is formed from a plurality of congruent first images of a training region of the body. Each first image is produced using an MRI pulse sequence different from the pulse sequences used to produce the other first images. Each first image comprise an array of pixels, and each training sample comprises a spatially aligned set of pixels from each first image. The same technique is used to produce a plurality of test samples corresponding to a test region of the same body. The test samples are produced using the same pulse sequences as the training samples. The training and test samples are then compared, to produce similarity data indicating, for each test sample, the degree of similarity between the test sample and the training samples. A display is then generated based upon the similarity data.

Dependencies are inferred from the claim text.