Systems and methods for accurate and rapid positron emission tomography using deep learning

1. A computer-implemented method for improving image quality with shortened acquisition time, comprising:

(a) acquiring, using a medical imaging apparatus, a medical image of a subject, wherein the medical image is acquired using an accelerated image acquisition parameter with a first image quality;

(b) applying a deep network model to the medical image to generate a corresponding transformed medical image with a second image quality, wherein the second image quality is improved over the first image quality in at least one of image resolution, signal to noise ratio, motion artifact, and shading; and

(c) combining at least physiological or biochemical information from the medical image with the corresponding transformed medical image to generate an output image, wherein the output image has a quantification accuracy enhanced over a quantification accuracy of the transformed medical image or an image quality enhanced over the first image quality of the medical image, wherein one or more parameters quantifying the quantification accuracy of the transformed medical image is selected from the group consisting of standardized uptake value (SUV), local peak value of SUV, maximum value of SUV, and mean value of SUV.

2. The computer-implemented method of claim 1 , wherein the medical image and the corresponding transformed medical image are dynamically combined based at least in part on a quantification accuracy of the medical image.

3. The computer-implemented method of claim 1 , wherein the medical image and the corresponding transformed medical image are spatially combined.

4. The computer-implemented method of claim 3 , wherein the medical image and the corresponding transformed medical image are combined using ensemble averaging.

5. The computer-implemented method of claim 1 , wherein the medical image and the corresponding transformed medical image are combined using an adaptive mixing algorithm.

6. The computer-implemented method of claim 5 , wherein the adaptive mixing algorithm comprises calculating a weighting coefficient for the medical image and the corresponding transformed medical image.

7. The computer-implemented method of claim 6 , wherein the weighting coefficient is calculated based on one or more parameters quantifying the quantification accuracy of the transformed medical image.

8. The computer-implemented method of claim 6 , wherein the weighting coefficient is calculated based on both an image quality and a quantification accuracy of the medical image and the corresponding transformed medical image.

9. The computer-implemented method of claim 1 , wherein the medical image is Positron Emission Tomography (PET) image.

10. A non-transitory computer-readable storage medium including instructions that, when executed by one or more processors, cause the one or more processors to perform operations comprising:

(a) acquiring, using a medical imaging apparatus, a medical image of a subject, wherein the medical image is acquired using an accelerated image acquisition parameter with a first image quality;

(b) applying a deep network model to the medical image to generate a corresponding transformed medical image with a second image quality, wherein the second image quality is improved over the first image quality in at least one of image resolution, signal to noise ratio, motion artifact, and shading; and

(c) combining at least physiological or biochemical information from the medical image with the corresponding transformed medical image to generate an output image, wherein the output image has a quantification accuracy enhanced over a quantification accuracy of the transformed medical image or an image quality enhanced over the first image quality of the medical image, wherein one or more parameters quantifying the quantification accuracy of the transformed medical image is selected from the group consisting of standardized uptake value (SUV), local peak value of SUV, maximum value of SUV, and mean value of SUV.

11. The non-transitory computer-readable storage medium of claim 10 , wherein the medical image and the corresponding transformed medical image are dynamically combined based at least in part on a quantification accuracy of the medical image.

12. The non-transitory computer-readable storage medium of claim 10 , wherein the medical image and the corresponding transformed medical image are spatially combined.

13. The non-transitory computer-readable storage medium of claim 12 , wherein the medical image and the corresponding transformed medical image are combined using ensemble averaging.

14. The non-transitory computer-readable storage medium of claim 10 , wherein the medical image and the corresponding transformed medical image are combined using an adaptive mixing algorithm.

15. The non-transitory computer-readable storage medium of claim 14 , wherein the adaptive mixing algorithm comprises calculating a weighting coefficient for the medical image and the corresponding transformed medical image.

16. The non-transitory computer-readable storage medium of claim 15 , wherein the weighting coefficient is calculated based on one or more parameters quantifying the quantification accuracy of the transformed medical image.

17. The non-transitory computer-readable storage medium of claim 15 , wherein the weighting coefficient is calculated based on both an image quality and a quantification accuracy of the medical image and the corresponding transformed medical image.

18. The non-transitory computer-readable storage medium of claim 10 , wherein the medical image is Positron Emission Tomography (PET) image.