System and method for non-invasive patient-image registration

摘要

A system for non-invasive registration between a patient and a three-dimensional (3D) medical image includes a near infrared 3D camera 110 which extracts a 3D blood vessel image I2 of a patients registration target area during surgical operation; a camera position tracer 120 which traces a position of the near infrared 3D camera 110 and calculates a real world coordinate system of the 3D blood vessel image I2; a controller 130 which extracts a first blood vessel pattern from a 3D medical image I1 of the registration target area, extracts a second blood vessel pattern from the 3D blood vessel image I2, and performs position registration between the patient and the 3D medical image I1 through the extracted first and second blood vessel patterns; and a display 140 which displays a registration result calculated by the controller 130.

主权项

1. A system for non-invasive registration between a patient and a three-dimensional (3D) medical image, the system comprising:
a near infrared 3D camera configured to extract a 3D blood vessel image of a patient's registration target area; a camera position tracer configured to trace a position of the near infrared 3D camera and to calculate a real world coordinate system of the 3D blood vessel image; a controller comprising:
a 3D medical image database configured to store a 3D medical image obtained by taking the registration target area;a blood vessel model generator configured to reconstitute a blood vessel model by separating a blood vessel image from the 3D medical image;a first feature point extractor configured to extract a branch point from which a blood vessel is branched by detecting a skeleton line of the blood vessel model, and to select each extracted branch point as a feature point of a first blood vessel pattern;a 3D blood vessel image detector configured to extract a 3D blood vessel image by separating a blood vessel image from a near infrared 3D image taken by the near infrared 3D camera, and a second feature point extractor configured to extract a branch point of a patient's blood vessel through the extracted 3D blood vessel image and a real world coordinate system of the 3D blood vessel image measured by the camera position tracer, and to select each extracted branch point as a feature point of a second blood vessel pattern; anda registrator configured to determine a registration result by optimally matching a first feature point group comprising of a plurality of first feature points, and a second feature point group comprising of a plurality of second feature points; and a display configured to display the registration result.