Image processing device, method, and program

Abstract

[Problem] To suppress misclassification of structures contained in an image. [Solution] The present invention is provided with: a filtering unit (32) that calculates a Hessian matrix by performing filtering on an image using a second order partial derivative of a solid spherical function; and an evaluation unit (30) that evaluates structures contained in the image using the eigenvector and eigenvalue of the Hessian matrix. The filtering unit (32) is provided with a correction unit (33) that calculates a first order partial derivative vector by performing filtering using the first order partial derivative of a solid spherical function having the same radius as the radius of a solid sphere in the image, and using the value resulting from projecting the first order partial derivative vector in the direction of the eigenvector, performs correction negating one response wave among response waves represented in two positions symmetrically separated from the center of the solid sphere of the second order partial derivative in each direction of the function representing the solid sphere.

Description

technical field [0001] The present invention relates to an image processing device and method for discriminating the structure of a structure in an image, and a program for causing a computer to execute the image processing method. Background technique [0002] In recent years, high-quality three-dimensional images have been used for image diagnosis with the advancement of medical equipment (for example, multi-detector CT, etc.). A three-dimensional image is composed of a plurality of two-dimensional images and has a large amount of information. Therefore, it may take time for a doctor to find a desired observation site and make a diagnosis. Therefore, by extracting the organ of interest and displaying MIP, VR, CPR, etc., the visibility of the whole organ or lesion can be improved and the efficiency of diagnosis can be improved. [0003] On the other hand, as a method of extracting blood vessels and bones in medical images, Hessian analysis using a Hessian matrix has been p...

AI Technical Summary

Problems solved by technology

A 3D image is composed of multiple 2D images and contains a lot of information, so it may take time for a doctor to find the desired observation site and make a diagnosis

[0004] However, in the method of Non-Patent Document 1, when there are other structures (peripheral structures) around the linear structure, a part of the surrounding structure may be misjudged as a linear structure.

Images