Variance reduction for monte carlo-based scatter estimation
A method rotates a radiation source and a detector over a sequence of acquisition angles about a subject and acquires, at each acquisition angle, a 2D projection image. A simulation subset is formed that contains some, but not all, of the acquired 2D projection images, wherein subset membership is determined according to relative signal change between successive 2D projection images. Scatter is characterized for the acquired sequence of projection images according to the formed simulation subset. One or more of the acquired sequence of projection images is corrected using the scatter characterization. An image volume is reconstructed and stored according to the sequence of scatter corrected projection images. One or more images of the reconstructed image volume are rendered to a display or transmitting the stored image volume data.
1. A method comprising:
DA accessing a sequence of 2D projection images of a subject acquired over a plurality of acquisition angles;
forming a simulation subset that includes some, but not all, of the 2D projection images, including:
determining the simulation subset according to an accumulated relative signal change between successive 2D projection images;
identifying one of the 2D projection images as a candidate projection image;
calculating a change value according to a relative signal change between the candidate projection image and an adjacent projection image immediately before or after the candidate projection image;
adding the calculated change value to an accumulated sum;
comparing the accumulated sum to a predetermined threshold value; and
if the threshold value is exceeded, assigning the candidate projection image to the simulation subset and incrementing the threshold value by a predetermined increment;
characterizing scatter for the sequence of projection images according to the formed simulation subset;
correcting one or more of the sequence of 2D projection images using the scatter characterization;
reconstructing and storing an image volume according to the sequence of scatter corrected projection images; and
rendering one or more images of the reconstructed image volume to a display or transmitting the stored image volume data.
2. The method of claim 1 , further comprising preprocessing the candidate projection image by scaling to reduced resolution.
3. The method of claim 1 , further comprising preprocessing the candidate projection image by converting the candidate projection image to log-exposure data form.
4. The method of claim 1 , further comprising smoothing the calculated change value.
5. The method of claim 4 , wherein smoothing is accomplished by applying Gaussian smoothing logic.
6. The method of claim 1 , further comprising identifying a simulation region within 2D projection images in the simulation subset.
7. The method of claim 1 , wherein characterizing scatter comprises performing a Monte Carlo-type simulation using the following adjustments:
increasing a first number of photons directed toward high density regions;
decreasing a second number of photons directed toward low density regions; and
adjusting photon weighting to reduce bias in scatter estimates.
8. The method of claim 1 , wherein the sequence of 2D projection images is acquired by rotating a radiation source and a detector over a sequence of acquisition angles about the subject to acquire, at each acquisition angle, a 2D projection image.
9. A method for scattered radiation characterization, comprising:
a) rotating a radiation source and a detector over a sequence of acquisition angles about a subject and acquiring, at each acquisition angle, a 2D projection image;
b) forming a simulation subset including some, but not all, of the acquired 2D projection images by successive iterations of a sequential selection process comprising:
(i) identifying one of the acquired 2D projection images as a candidate projection image;
(ii) calculating a change value according to a relative signal change between the candidate projection image and an adjacent 2D projection image acquired immediately before or after the candidate projection image;
(iii) adding the calculated change value to an accumulated sum;
(iv) comparing the accumulated sum to a predetermined threshold value; and
(v) if the threshold value is exceeded, assigning the candidate projection image to the simulation subset; and updating the threshold value;
c) characterizing scatter for the acquired sequence of projection images according to the formed simulation subset;
d) correcting one or more of the acquired sequence of 2D projection images using the scatter characterization;
e) reconstructing and storing an image volume according to the sequence of scatter corrected projection images; and
f) rendering one or more images of the reconstructed image volume to a display or transmitting the stored image volume data.
10. The method of claim 9 , wherein characterizing scatter comprises performing a Monte Carlo-type simulation.
11. The method of claim 9 , further comprising preprocessing the candidate projection image by scaling to reduced resolution.
12. The method of claim 9 , further comprising preprocessing the candidate projection image by converting the candidate projection image to log-exposure data form.
13. The method of claim 9 , further comprising smoothing the calculated change value.
14. The method of claim 13 , wherein smoothing is accomplished by applying Gaussian smoothing logic.
15. The method of claim 9 , further comprising identifying a simulation region within 2D projection images in the simulation subset.
16. The method of claim 9 , wherein characterizing scatter comprises performing a Monte Carlo-type simulation using the following adjustments:
increasing a first number of photons directed toward high density regions;
decreasing a second number of photons directed toward low density regions; and
adjusting photon weighting to reduce bias in scatter estimates.
17. The method of claim 9 , wherein the sequence of 2D projection images is acquired by rotating a radiation source and a detector over a sequence of acquisition angles about the subject to acquire, at each acquisition angle, a 2D projection image.
18. A method for scattered radiation characterization, the method comprising:
receiving a sequence of 2D projection images of a subject acquired over a plurality of projection angles;
forming a simulation subset that includes some, but not all, of the received 2D projection images based on measured signal level changes as between successively acquired ones of the 2D projection images;
for each 2D projection image in the simulation subset, defining a simulation region on the radiation detector for x-ray scatter estimation, including identifying a region on the radiation detector as a low exposure region based on exposure values of the low exposure region satisfying a preset exposure threshold and defining the simulation region as the low exposure region;
characterizing x-ray scatter for the received sequence of 2D projection images in the simulation subset using simulated x-ray scatter impinging the defined simulation region and ignoring simulated x-ray scatter impinging areas outside the defined simulation region;
correcting the received sequence of 2D projection images using the characterized scatter; and
reconstructing a 3D image volume according to the sequence of scatter corrected projection images.
19. The method of claim 18 , wherein the step of characterizing x-ray scatter comprises executing Monte Carlo simulation for photon energy incident on the defined simulation region and ignoring photon energy directed outside the simulation region.
20. The method of claim 18 , further comprising receiving a sequence of 2D projection images of a human hand, elbow, knee, or ankle acquired over the plurality of acquisition angles.