Medical device and method for processing microwave medical images with 3D shape and texture feature descriptors to characterize lesions in imaged tissue.

Abstract

The invention relates to a method for processing medical images of a patient's body zone, in particular human tissue of the breast, by means of a microwave medical imaging device, the method comprising the following steps implemented in a processing unit of the medical imaging device: - identifying at least one region of interest using at least one initial microwave image of the patient's body zone; - processing each region of interest identified in the image to determine at least a first shape feature, preferably a solidity of each region of interest, and at least a second and a third feature related to the texture of each region of interest, the first, second and third features being coordinates characterizing each region of interest; - locating each region of interest based on its coordinates in a space of at least three dimensions, these dimensions being at least the first, second and third features, respectively, said space being divided by a decision hypersurface into two continuous and distinct subspaces, one subspace in which regions of interest located are associated with benign lesions and another subspace in which regions of interest located with malignant lesions;

Claims

1. A method for processing microwave medical images of human tissue, particularly the breast, performed by a computer, - A step of identifying at least one region of interest from at least one initial microwave image of a zone of the patient's body, - A step of processing each region of interest identified in the image in order to determine only the 3 to 5 features relating to the aforementioned region of interest, wherein the features are - At least a first shape feature, including one relating to the solidity of each region of interest, - At least two second and third features relating to the texture of each region of interest, two of which are correlation and busyness, The first, second, and third features are the coordinates that characterize each region of interest, namely the step, A step of positioning each region of interest based on its coordinates in a space of at least three dimensions, wherein each dimension has at least a first, second, and third feature, and the space is divided into two continuous and distinct subspaces by a determined hypersurface, one subspace such that the region of interest located therein is associated with a benign lesion, and the other subspace such that the region of interest located therein is associated with a malignant lesion. A method that includes this.

2. The method according to claim 1, comprising the step of determining a classification score for each region of interest, wherein the score corresponding to the probability of malignant lesions is the posterior probability of the region of interest belonging to the class of malignant lesions.

3. The method according to claim 2, wherein if the probability of the malignant lesion is 50% or less, the region of interest is associated with a benign lesion, and if the probability of the malignant lesion is greater than 50%, the region of interest is associated with a malignant lesion.

4. The method according to claim 1, wherein the second feature is a measurement of the spatial relationship between the intensities of pixels in a region of interest in a specific direction.

5. The method according to claim 1, wherein the third feature is a measurement of the spatial relationship between the intensities of three or more adjacent pixel groups in the region of interest.

6. The method according to claim 1, wherein the decision surface is obtained by a simple Bayesian classifier pre-trained for training the region of interest or by a quadratic discriminant analysis classifier.

7. The method according to claim 1, further comprising a processing step for each region of interest to refine the contour of each region of interest.

8. The method according to claim 1, comprising the steps of obtaining at least one initial image of a zone to be imaged, and performing morphological processing on each initial image to identify a region of interest.

9. The method according to claim 1, comprising the step of verifying the region of interest by evaluating its persistence across multiple morphological images.

10. A medical image processing apparatus comprising a processing unit configured to carry out the method described in claim 1.

Images