Machine learning architecture for imaging protocol detector

1. A system comprising:

a capture device configured to capture a first image representing at least a portion of a mouth of a user;

a communication device configured to communicate user feedback to the user using a display; and

a processor and a non-transitory computer-readable medium containing instructions that when executed by the processor causes the processor to perform operations comprising:

receiving the first image representing at least the portion of the mouth of the user; and

outputting user feedback for capturing a second image representing at least a portion of the mouth of the user, wherein the user feedback is output in response to using a machine learning architecture to determine that an image quality score of the first image does not satisfy an image quality threshold, the user feedback comprising matching, using the display, a first object overlaid on top of the first image representing at least the portion of the mouth of the user to a second object overlaid on top of the first image representing at least the portion of the mouth of the user.

2. The system of claim 1 , wherein the capture device is further configured to stream a video, the video comprising a plurality of frames, at least one of the frames representing the portion of the mouth of the user.

3. The system of claim 2 , wherein the processor is further configured to perform operations comprising automatically identifying a frame of the plurality of frames as a high quality image, the frame being identified as the high quality image in response to determining via the machine learning architecture that the image quality score of the frame satisfies the image quality threshold.

4. The system of claim 1 , wherein the communication device is configured to communicate user feedback to the user using audio.

5. The system of claim 1 , wherein the communication device is configured to communicate user feedback to the user using haptic feedback.

6. The system of claim 1 , wherein the capture device comprises a light, the capture device configured to activate the light in response to the user feedback.

7. A computer-implemented method comprising:

receiving, by a machine learning architecture executing on one or more processors, a first image representing at least a portion of a mouth of a user;

determining, by the machine learning architecture, user feedback for outputting to the user, wherein the user feedback is determined based on a machine learning model of the machine learning architecture, the machine learning model trained by calculating a reward in response to an action and a policy; and

outputting, by the machine learning architecture, the user feedback, the user feedback for capturing a second image representing a portion of the mouth of the user, wherein the machine learning architecture outputs the user feedback in response to an image quality score of the first image not satisfying an image quality threshold.

8. The computer-implemented method of claim 7 , wherein the image quality score comprises at least one of an image quality score with respect to an image characteristic or an image quality score with respect to image content.

9. The computer-implemented method of claim 8 , wherein the image quality score with respect to the image characteristic is determined using a machine learning model of the machine learning architecture, the machine learning model trained using a training dataset comprising a distorted image and a corresponding clean image.

10. The computer-implemented method of claim 8 , wherein the image quality score with respect to image content is determined using a machine learning model of the machine learning architecture, the machine learning model trained using a class label to identify a class in a training image.

11. The computer-implemented method of claim 7 , wherein the image quality score is used to predict at least one of a future image characteristic or a future image content.

12. The computer-implemented method of claim 7 , further comprising:

selecting, by the machine learning architecture, a user feedback phrase, the user feedback phrase being output to the user as the user feedback, the user feedback phrase selected in response to a user feedback type output from the machine learning model, the user feedback phrase selected from a dictionary of phrases.

13. The computer-implemented method of claim 7 , further comprising:

receiving, by the one or more processors, a video comprising a plurality of frames, at least one of the frames representing the portion of the mouth of the user; and

extracting, by the one or more processors, the at least one frame representing the portion of the mouth of the user as the first image.

14. The computer-implemented method of claim 7 , further comprising:

receiving, by the machine learning architecture, the second image representing at least the mouth of the user;

causing, by the machine learning architecture, the second image to be provided to a server in response to an image quality score of the second image satisfying the image quality threshold.

15. The computer-implemented method of claim 7 , wherein the machine learning architecture executes on one or more processors of a device, and wherein the first image is captured by the device.

16. The computer-implemented method of claim 7 , wherein the machine learning architecture executes on one or more processors of a server, and wherein the first image is captured by a device of the user.

17. The computer-implemented method of claim 7 , further comprising:

receiving, by the machine learning architecture, another image representing at least a portion of the mouth of the user; and

outputting, by the machine learning architecture, user feedback in response to an image quality score of the another image satisfying the image quality threshold.

18. The computer-implemented method of claim 7 , further comprising:

receiving, by the machine learning architecture, a third image; and

outputting, by the machine learning architecture, user feedback in response to determining via the machine learning architecture that the third image does not represent at least a portion of the mouth of the user.

19. A system comprising:

a communication device configured to capture a first image representing at least a portion of a mouth of a user and communicate the first image to a server; and

a processor of the server and a non-transitory computer-readable medium containing instructions that when executed by the processor causes the processor to perform operations comprising:

receiving the first image representing at least the portion of the mouth of the user;

selecting, using a machine learning architecture, a user feedback phrase, the user feedback phrase selected in response to a user feedback type output from a machine learning model, the user feedback phrase selected from a dictionary of phrases; and

communicating, to the communication device, user feedback for capturing a second image representing at least a portion of the mouth of the user, wherein the user feedback is determined in response to determining via an imaging protocol algorithm that an image quality score of a portion of the first image does not satisfy an image quality threshold, wherein the user feedback comprises the user feedback phrase.

20. The system of claim 19 , wherein the instructions when executed by the processor further cause the processor to transmit a command to the communication device using an application programming interface.

21. The system of claim 19 , wherein the server communicates to the communication device using an application on the communication device supported by a corresponding application on the server.

22. The system of claim 19 , wherein the instructions when executed by the processor further cause the processor to receive a second portion of the first image satisfying an image quality threshold.

23. The system of claim 19 , wherein the instructions when executed by the processor further cause the processor to bias the user feedback to improve the portion of the first image that does not satisfy the image quality threshold.