Systems, devices and methods for non-invasive hematological measurements

1. A system, comprising:

a platform to receive a body portion of a user during use;

an imaging device coupled to the platform and to acquire a set of images of at least a capillary bed of the body portion;

a processor, communicably coupled to the imaging device, to:

detect, by the processor, in each image of the set of images, one or more capillaries in the capillary bed of the body portion to identify a first set of capillaries across the set of images, including:

estimating one or more attributes of each capillary of the first set of capillaries; and

when a first attribute of the one or more attributes of each capillary of the first set of capillaries meets a predetermined criterion for the first attribute, selecting that capillary for inclusion in the first set of capillaries, the one or more attributes including one or more structural attributes, one or more flow attributes, one or more imaging attributes, or combinations thereof; and

identify a subset of capillaries from the first set of capillaries as a second set of capillaries such that each capillary of the second set of capillaries is visible in at least a predetermined number of images of the set of images.

2. The system of claim 1 , wherein the imaging device acquires the set of images as a set of frames of a video.

3. The system of claim 1 , further comprising an illumination source coupled to the platform and to illuminate the body portion, wherein the imaging device acquires the set of images in response to the illumination of the body portion.

4. The system of claim 1 , the one or more structural attributes selected from the group consisting of average capillary diameter, lateral capillary diameter, vertical capillary diameter capillary length, and capillary shape.

5. The system of claim 1 , the one or more imaging attributes selected from the group consisting of contrast, focus, signal-to-noise ratio, and image stability.

6. The system of claim 1 , wherein the first attribute is average capillary diameter, wherein each capillary of the second set of capillaries has an estimated average capillary diameter from about 10 μm to about 20 μm.

7. The system of claim 1 , wherein the processor further:

detects, for the set of images and in the second set of capillaries, a set of cellular events, each cellular event of the set of cellular events associated with passage of a white blood cell in a capillary of the second set of capillaries; and

estimates an event count for the second set of capillaries based on the set of cellular events.

8. The system of claim 7 , wherein the processor further:

for each capillary of the second set of capillaries, estimates a quality factor;

estimates the event count based on the set of cellular events and the quality factor associated with each capillary of the second set of capillaries.

9. The system of claim 8 , wherein the processor further:

receives a set of training images associated with capillary beds in body portions of a set of training users;

generates, via supervised learning, an event count threshold based on the set of training images;

classifies the user to a first user type of a set of user types based on the event count and the event count threshold, at least one user type of the set of user types associated with a diagnosis of neutropenia; and

transmits an indication of the first user type to the user.

10. The system of claim 1 , wherein the processor further generates a confidence value associated with the image of each capillary of the first set of capillaries in the set of images, the first set of capillaries including those capillaries for which the confidence value, for each image in which that capillary is detected, exceeds a confidence threshold.

11. The system of claim 1 , wherein the processor detects the first set of capillaries by:

receiving a set of training images including a specification of one or more capillaries visible within each image of the set of training images;

training a neural network on the set of training images; and

applying the set of images to the neutral network to detect the first set of capillaries.

12. The system of claim 1 , wherein the processor detects the first set of capillaries by applying the set of images to a neutral network, the neural network being trained on a set of training images including a specification of one or more capillaries visible within each image of the set of training images.

13. A method, comprising:

acquiring a set of images of a capillary bed of a body portion of a user;

detecting, without user input, in each image of the set of images, one or more capillaries in the body portion to identify a first set of capillaries across the set of images, including:

estimating one or more attributes of each capillary of the first set of capillaries; and

when a first attribute of the one or more attributes of each capillary of the first set of capillaries meets a predetermined criterion for the first attribute, selecting that capillary selected for inclusion in the first set of capillaries, the one or more attributes including one or more structural attributes, one or more flow attributes, one or more imaging attributes, or combinations thereof; and

identifying a subset of capillaries from the first set of capillaries as a second set from the first set of capillaries such that each capillary of the second set of capillaries is visible in at least a predetermined number of images of the set of images.

14. The method of claim 13 , the acquiring including acquiring the set of images as a set of frames of a video.

15. The method of claim 13 , further comprising illuminating the body portion, the acquiring the set of images being in response to the illumination of the body portion.

16. The method of claim 13 , the one or more structural attributes selected from the group consisting of average capillary diameter, lateral capillary diameter, vertical capillary diameter capillary length, and capillary shape.

17. The method of claim 13 , the one or more imaging attributes selected from the group consisting of contrast, focus, signal-to-noise ratio, and image stability.

18. The method of claim 13 , wherein the first attribute is average capillary diameter, wherein each capillary of the second set of capillaries has an estimated average capillary diameter from about 10 μm to about 20 μm.

19. The method of claim 13 , further comprising:

detecting, for the set of images and in the second set of capillaries, a set of cellular events, each cellular event of the set of cellular events associated with passage of a white blood cell in a capillary of the second set of capillaries; and

estimating an event count for the second set of capillaries based on the set of cellular events.

20. The method of claim 19 , further comprising:

for each capillary of the second set of capillaries, estimating a quality factor;

estimating the event count based on the set of cellular events and the quality factor associated with each capillary of the second set of capillaries.

21. The method of claim 20 , further comprising:

receiving a set of training images associated with capillary beds in body portions of a set of training users;

generating, via supervised learning, an event count threshold based on the set of training images; and

classifying the user to a first user type of a set of user types based on the event count and the event count threshold, at least one user type of the set of user types associated with a diagnosis of neutropenia; and

transmitting an indication of the first user type to the user.

22. The method of claim 13 , further comprising generating a confidence value associated with the image of each capillary of the first set of capillaries in the set of images, the first set of capillaries including those capillaries for which the confidence value, for each image in which that capillary is detected, exceeds a confidence threshold.

23. The method of claim 13 , the detecting the first set of capillaries further including:

receiving a set of training images including a specification of one or more capillaries visible within each image of the set of training images;

training a neural network on the set of training images; and

applying the set of images to the neutral network to detect the first set of capillaries.

24. The method of claim 13 , the detecting the first set of capillaries further including applying the set of images to a neutral network, the neural network being trained on a set of training images including a specification of one or more capillaries visible within each image of the set of training images.

25. The method of claim 13 , wherein the body portion is a nailfold portion of a finger of the user.

26. A device comprising a processor to:

receive a set of images of a capillary bed in a body portion of a user;

detect, by the processor, in each image of the set of images, one or more capillaries in the body portion to identify a first set of capillaries across the set of images, including:

estimating one or more attributes of each capillary of the first set of capillaries: and

when a first attribute of the one or more attributes of each capillary of the first set of capillaries meets a predetermined criterion for the first attribute, selecting that capillary for inclusion in the first set of capillaries, the one or more attributes including one or more structural attributes, one or more flow attributes, one or more imaging attributes, or combinations thereof, such that a first attribute of the one or more attributes of each capillary of the first set of capillaries meets a predetermined criterion for the first attribute;

identify a subset of capillaries from the first set of capillaries as a second set of capillaries such that each capillary of the second set of capillaries is visible in at least a predetermined number of images of the set of images;

detect, for the set of images and in the second set of capillaries, a set of cellular events, each cellular event of the set of cellular events associated with passage of a white blood cell in a capillary of the second set of capillaries; and

estimate an event count for the second set of capillaries based on the set of cellular events.

27. The device of claim 26 , wherein the processor further:

for each capillary of the second set of capillaries, estimates a quality factor;

estimates the event count based on the set of cellular events and the quality factor associated with each capillary of the second set of capillaries.

28. The device of claim 27 , wherein the processor further:

receives a set of training images associated with capillary beds in body portions of a set of training users;

generates, via supervised learning, an event count threshold based on the set of training images; and

classifies the user to a first user type of a set of user types based on the event count and the event count threshold, at least one user type of the set of user types associated with a diagnosis of neutropenia; and

transmits an indication of the first user type to the user.

29. The device of claim 26 , wherein the processor further generates a confidence value associated with the image of each capillary of the first set of capillaries in the set of images, the first set of capillaries including those capillaries for which the confidence value, for each image in which that capillary is detected, exceeds a confidence threshold.

30. The device of claim 26 , wherein the processor detects the first set of capillaries by:

receiving a set of training images including a specification of one or more capillaries visible within each image of the set of training images;

training a neural network on the set of training images; and

applying the set of images to the neutral network to detect the first set of capillaries.

31. The device of claim 26 , wherein the processor detects the first set of capillaries by applying the set of images to a neutral network, the neural network being trained on a set of training images including a specification of one or more capillaries visible within each image of the set of training images.