Automatic interpretation of 3-D medicine images of the brain and methods for producing intermediate results
AssigneeExini Diagnostics Aktiebolag
InventorsDavid Jakobsson, Jens Richter, Andreas Järund
Life of this patent
Estimating the term…
Abstract
Methods for fully automatic quantification and interpretation of three dimensional images of the brain or other organs. A system for Computer Aided Diagnosis (CAD) of diseases affecting cerebral cortex from SPECT images of the brain, where said images may represent cerebral blood flow (CBF). The methods include image processing, statistical shape models, a virtual brain atlas, reference databases and machine learning.
Claims10
A method for automatic interpretation of a three dimensional medicine image representing a body organ where said method comprises the following steps:
- (a) automatically rotating the images to adjust for accidental tilting when the images were originally taken;
- (b) automatically finding the contours of the organ;
- (c) automatically adjusting size, position, rotation, and shape of a predefined model shape of the type of organ in question to fit the organ of the current image;
- (d) automatically, with the aid of the model shape, defining certain portions of the image;
- (e) normalizing the intensity of the image;
- (f) quantifying each point in the image of the organ, producing a quantification result;
- (g) feeding the quantification results to an interpretation system;
- (h) letting the interpretation system interpret the image, producing an interpretation result;
- and
- (i) presenting the interpretation result.
A method for automatic normalization of brain images comprising the step of:
- using a brain atlas and active shape models with a cerebellum as reference.
A method for reorientation of brain images into a common coordinate system, comprising the following steps:
- for each reference brain, providing a midline sagittal slice;
- for the midline sagittal slice, finding and marking an occipital-cerebellum notch;
- for the midline sagittal slice, finding and marking the lowest point of the frontal lobe;
- drawing a line between the occipital-cerebellum notch and the lowest point of the frontal lobe;
- orienting the image such that said line becomes horizontal, and parallel to the x-axis;
- finding foremost point of frontal lobe;
- position the image such that the x-axis runs through foremost point of frontal lobe;
- finding the cerebellum-temporal lobe notch;
- and
- positioning the image such that the y-axis runs through said cerebellum-temporal lobe notch.
A method, comprising:
- determining a contour of an organ in a patient's body, from an image of the organ in at least three dimensions, by adjusting a predefined contour model to the image, wherein said body organ is a brain;
- wherein determining a contour of the brain is performed by an image contour modeling system configured for modeling the image of the brain into one of two or more classes where one class is normal and one class is pathological, the image modeling system comprising:
- a pretrained artificial neural network having a plurality of input nodes, and a number of output nodes;
- and
- a feature extractor, capable of extracting a number of features from the image, said features being suitable for feeding to the input nodes;
- wherein the pretrained artificial network presents a classification result on the number of output nodes when the number of features of the image is fed to the plurality of input nodes.
Dependencies are inferred from the claim text.