US 2013/0259339 A1Published application

Tomography system based on cerenkov luminescence

AssigneeInstitute of Automation, Chinese Academy of Sciences

InventorsJie Tian, Jianghong Zhong, Xin Yang, Chenghu Qin

Life of this patent

Estimating the term…

Loading patent…

Abstract

A tomography system based on Cerenkov tomography, comprising: a detector of Cerenkov fluorescence for acquiring optical plane images; a structural imaging system for acquiring three-dimensional structural images; a bed device for supporting an object to be imaged; a computer for forming an optical image, a structural image and a CLT image. The invention adopts the SP3 model and the semi threshold iterator to implement the global reconstruction of the CLT, and obtains the three-dimensional tomography to image of the distribution of the radiopharmaceutical and the molecular probe in vivo within a short time. Since ordinary CCD camera is used, the cost of the imaging system has been sharply reduced for the equipment's construct and maintenance compared with PET/SPECT or γ camera. Therefore the present invention expands the options of the molecular probe, and application of the is medicine Imaging.

A tomography method based on Cerenkov Effect, comprises:

Acquiring optical plane images and three-dimensional structural images;
obtaining a space registration of the optical plane image acquired by different perspective and the three-dimensional structural image based on fixed marking points, and constructing a geometrical mapping table of an optical and a tomography image;
Converting every pixel's gray value of the optical plane image into an illumination intensity of three-dimensional object based on the geometrical mapping table, and obtaining a distribution B of the illumination intensity on a surface of the object;
Segmenting the three-dimensional tomography image In order to get each organ, and the organs being assigned by different pixel values;
The segmented body data being discretized by a finite element method, and outputting a volume gridding;
Based on the SP3 forward model, constructing a linear mapping table between the distribution of the surface light intensity B and the unknown distribution of the isotope in vivo S, B=AS, where A stands for the system parameter matrix of the forward model, B stands for a m-dimensional column vector, S stands for an n-dimensional column vector. m and n are both signless integral, and m<n;
Based on the theory of the L1/2 regularization, objective function F(S) being calculated for reconstructing tomography image using a iterator as follows:
F(S)=min {∥B−AS∥2+p∥s∥1/2}, where p stands for the parameter of the regularization, p=(1−ε)/∥A∥2
In the K+1st iterate, the iterator update the distribution of the isotope in vivo. Sk+1=Hpq(Tk(Sk))Tk(Sk)=SkqkAT(B−ASk), where k is a signless integral, qk=4|Tk|3/2∥A∥2/3, H=(h(S1), h(S2) . . . h(Sn)T,
h

(
S
)
=
2

S

(
1
+
cos

2

π
-
2

φ
3
)
/
3
,
when S>η(pq)3/2, h(S)=0 and S≦η(pq)3/2,
φ
=
ar


cos
(
0.125


pq

(

S

3
)
-
3
/
2
)
,
η stands for a real constant;
If the S got by the k st makes the objective function Fk<V, stop the iterator, where V is a threshold constant.

60 more elements

Abstract

Claims10