De-noised reconstructed image data edge improvement

1. A method, comprising:

obtaining de-noised reconstructed tomographic image data;

generating noisy edge image data by multiplying noise image data and an edge map, wherein the noise image data represents noise removed from reconstructed tomographic image data used to generate the de-noised reconstructed tomographic image data;

weighting the noisy edge image data with a predetermined weight by multiplying the noisy edge image data by the weight; and

adding the weighted noisy edge image data to the de-noised reconstructed tomographic image data, which edge improves a sub-set of the de-noised reconstructed tomographic image data corresponding to edges of structure represented in the de-noised reconstructed tomographic image data.

2. The method of claim 1 , wherein the edges are improved through noise cancellation.

3. The method of claim 1 , further comprising:

adding a negative of the noise image data to the de-noised reconstructed tomographic image data, wherein the noise image data is generated by subtracting the de-noised reconstructed tomographic image data from the reconstructed tomographic image data.

4. The method of claim 1 , wherein the weight has a value between zero and one.

5. The method of claim 1 , further comprising:

determining the weight based on a smoothness criterion.

6. The method of claim 5 , wherein the smoothness criterion is one of a total variation of an image or a squared difference of neighboring pixels of an image.

7. The method of claim 6 , further comprising:

determining the weight by minimizing a cost function, which is the smoothness criterion applied on the sum of the de-noised reconstructed tomographic image data and the weighted noisy edge image data.

8. The method of claim 1 , further comprising:

generating the edge map by applying an edge detection algorithm to the de-noised reconstructed tomographic image data.

9. The method of claim 8 , wherein the edge detection algorithm includes one or more of a weighted gradient magnitude, Canny or a watershed transform.

10. The method of claim 1 , wherein at least one of the edge map represents edge locations within the de-noised reconstructed tomographic image or the edge locations are located through values decreasing from a value of one to a value of zero within a predetermined window about an edge and zeroes everywhere else.

11. The method of claim 10 , wherein the edge locations are located through values of one at the edges and zeroes everywhere else.

12. The method of claim 1 , further comprising, at least one of:

dilating the edge map prior to the multiplication, smoothing the noise image data with a filter prior to the multiplication, or generating the noise image data by subtracting the de-noised reconstructed tomographic image data from the noisy reconstructed tomographic image data.

13. A system, comprising:

an edge detector, implemented by a processor, that detects an edge map of edge locations within de-noised reconstructed tomographic image data;

a noise image data generator, implemented by the processor, that generates noise image data by subtracting the reconstructed tomographic image data by the de-noised reconstructed tomographic image data;

a noisy edge image data generator, implemented by the processor, that generates noisy edge image data by multiplying the noise image data and the edge map; and

an edge improver, implemented by the processor, that generates edge improved de-noised image data by adding the noisy edge image data and a product of a weight and the de-noised reconstructed tomographic image data.

14. The system of claim 13 , wherein the edge improver cancels noise from the edges of structure in the de-noised reconstructed tomographic image data.

15. The system of claim 13 , wherein the weight is a value between negative one and zero.

16. The system of claim 13 , wherein the edge map represents one of edge locations within the de-noised reconstructed tomographic image using values decreasing from a value of one to a value of zero within a predetermined window about each edge and zeroes everywhere else or edge locations within the de-noised reconstructed tomographic image using a value of one at each edge and zeroes everywhere else.

17. The system of claim 13 , the edge improver, comprising:

a weight determiner, implemented by the processor, that determines the weight based on a smoothness criterion, which includes one of a total variation of an image or a squared difference of neighboring pixels of an image, by minimizing a cost function of the smoothness criterion being applied on the sum of the de-noised reconstructed tomographic image data and the weighted noisy edge image data.

18. A computer readable storage medium encoded with one or more computer executable instructions, which, when executed by a processor of a computing system, causes the processor to:

obtain reconstructed tomographic image data;

obtain de-noised reconstructed tomographic image data, wherein the de-noised reconstructed tomographic image data is a de-noised version of the obtained reconstructed tomographic image data;

generate an edge map of edge locations within the de-noised reconstructed tomographic image data by applying an edge detection algorithm to the de-noised reconstructed tomographic image data;

generate noise image data by subtracting the reconstructed tomographic image data by the de-noised reconstructed tomographic image data;

generate noisy edge image data by multiplying the noise image data and the edge map; and

generate edge improved de-noised image data by adding the noisy edge image data and a product of a weight and the de-noised reconstructed tomographic image data.