System and method of variable filter length local tomography

摘要

Methods, processes and systems of image reconstruction using variable filter length local tomography, for reconstructing internal body images in medical applications, and the like. The system and method of the present invention utilizes less radiation and less computer power than the prior art, without using iteration algorithms so that all target sizes from large to small can be reconstructed.

主权项

1. A method for reconstructing an image from cone beam (CB) projection data provided by at least one detector, the method comprising:
scanning an object to collect cone-beam (CB) projection data; loading the CB projection data into a computer; reconstructing an image of the object using the CB projection data; and repeatedly loading and using additional CB projection data until image reconstruction is finished at all reconstruction points x;
wherein reconstructing an image of the object using the CB projection data comprises only using truncated CB data at one or more reconstruction points x inside the object,wherein the truncated CB data is defined as data which is insufficient for theoretically exact reconstruction at a point x;wherein reconstructing an image of the object using the CB data further comprises:
filtering the truncated CB data; andback-projection updating of the image being reconstructed;wherein filtering the truncated CB data uses a filter of a length longer than the length of a filter associated with computing a derivative, but shorter than the length of a filter associated with theoretically exact reconstruction (of infinite length or of length equal to a support of the object along the line of filtration);wherein the image reconstructed by repeatedly loading and using additional CB projection data is not intended to approximate a theoretically exact reconstruction, but allows differentiation of tissue types that comprise the object being scanned, andwherein the computer does not process the projection data for the image reconstruction using iterative steps, does not perform filling in of the missing data, and does not use wavelets and curvelets.