System and method for assisted patient positioning
A method of obtaining a medical image includes obtaining, via a camera, at least one surface image of a patient. A pose of the patient is determined from the at least one surface image of the patient using at least one spatial information module. The patient is positioned, via a moveable bed, to an imaging start position and a medical image of the patient is obtained using a medical imaging modality.
1. A method for generating a medical image, comprising:
obtaining, via a camera, at least one surface image of a patient;
determining, from the at least one surface image, a pose of the patient using at least one spatial information module, the at least one spatial information module comprising a Dense Block Hourglass (DBHG) network;
positioning, via a moveable bed, the patient at an imaging start position; and
obtaining, via a medical image modality, a medical image of the patient.
2. The method of claim 1 , wherein the at least one spatial information module comprises at least one of a single-frame analytics network or a multi-frame analytics network.
3. The method of claim 2 , wherein the single-frame analytics network comprises a fully convolutional neural network.
4. The method of claim 1 , comprising preprocessing the at least one surface image of the patient prior to determining a pose of the patient.
5. A method for generating a medical image, comprising:
obtaining, via a camera, at least one surface image of a patient;
determining, from the at least one surface image, a pose of the patient using at least one spatial information module, the at least one spatial information module comprising a first neural network;
positioning, via a moveable bed, the patient at an imaging start position; and
obtaining, via a medical image modality, a medical image of the patient
generating a composite image from two or more of the surface images of the patient from different times, wherein the composite image is generated by a temporal information module in signal communication with the spatial information module.
6. The method of claim 5 , wherein the temporal information module comprises a relative motion recovery module and an image synthesis module.
7. The method of claim 6 , wherein the relative motion recovery module comprises a synchronized patient bed position reading to infer a relative motion between the patient and the camera.
8. The method of claim 6 , wherein the relative motion recovery module comprises a structure from motion module, a simultaneous localization and mapping module, or a combination thereof.
9. The method of claim 6 , wherein the image synthesis module comprises a depth fusion module.
10. The method of claim 9 , wherein the depth fusion method aggregates the depth map for each frame estimated from multi-view stereo matching process.
11. The method of claim 9 , wherein the image synthesis module synthesizes an extended field of view based on a coarse geometry fitting to the patient.
12. The method of claim 5 , wherein the temporal information module comprises a recurrent neural network.
13. The method of claim 5 , wherein the at least one spatial information module comprises a Dense Block Hourglass (DBHG) network.
14. A system of generating a medical image, comprising:
an imaging modality configured to obtain a medical image of a patient;
an imaging device configured to obtain at least one surface image of a patient; and
a processor configured to implement a combination of a spatial information module and a temporal information module, wherein the temporal information module is configured to generate a composite image from two or more surface images of the patient from different times, wherein the processor is configured to receive the composite image of the patient and verify at least one of a patient pose or a body region, and wherein the processor is configured to execute a medical imaging procedure using the imaging modality when the patient pose or body region is verified.
15. The system of claim 14 , wherein the spatial information module comprises a fully convolutional neural network.
16. The system of claim 14 , wherein the spatial information module comprises a Dense Block Hourglass (DBHG) network.
17. The system of claim 14 wherein the temporal information module comprises a relative motion recovery module and an image synthesis module.
18. The system of claim 17 , wherein the relative motion recovery module comprises a simultaneous localization and mapping module, a synchronized bed position module configured to infer a relative motion between the patient and the camera, or a combination thereof.
19. A non-transitory computer-readable medium encoded with computer executable instructions, the computer executable instructions, when executed by a computer in a system for reviewing image scans, cause the system for reviewing image scans to execute the steps of:
receiving at least one surface image of a patient;
determining a pose of the patient from the at least one surface image using at least one spatial information module comprising a machine-learned neural network having stacked first and second encoder-decoder pairs with the second encoder-decoder pair arranged to receive an input and output of the first encoder-decoder pair;
positioning, via a moveable bed, the patient at an imaging start position; and
obtaining, via a medical image modality, a medical image of the patient.