Method of measuring animal health using motion vector analysis
A method of measuring animal health, such as rodent health, such as a degree of arthritis, is described. Steps include capturing video frames, then using optical flow to compute vector fields for each frame, then taking a maximum vector value over a first time period, then selecting a set of maximum vector values over a second time period, then clipping those values to a maximum, then computing an average of those values. This is an animal health metric. Metrics may be organized into a plot over time and may then be compared against known plots to compute an overall animal health and measure efficacy of treatments or other animal characteristics or behavior, including predictive.
1. A method for measuring an animal health comprising the steps:
(a) recording, using a camera, a series of image frames of an animal cage interior wherein the cage comprises the animal enrolled in a medical study;
(b) generating a reduced vector field from at least a subset of image frames from step (a) using optical flow analysis;
(c) selecting one or more maximum vector values from the vector field from step (b);
(d) computing a first-period maximum value for the one or more maximum vector values from step (c) over a first time period length;
(e) repeating steps (a) through (d) for multiple, sequential time periods having the first time period length, wherein the multiple, sequential time periods fall within a second time period;
(f) selecting a second-period maximum value set from all of the first-period maximum values computed from repeating step (d);
(g) computing a mean value from the values of the second-period maximum value set selected from step (f);
(h) computing a selected animal health metric from the set: {degree of arthritis, arthritis index, efficacy of a drug; efficacy of a disease treatment, probability likelihood of having a disease; expected time to die} by subtracting a baseline value from the mean value from step (g); and
(i) plotting the computed animal health metric from step (h) over time.
2. The method for measuring an animal health of claim 1 comprising the additional step:
(j) masking each image in the series of image frames.
3. The method for measuring an animal health of claim 1 comprising the additional step:
(k) clipping the values in the maximum value set from step (f) to a predetermined maximum value, prior to computing a mean in step (g).
4. The method for measuring an animal health of claim 1 comprising the additional step of:
(m) comparing the plot from step (1) to a set of reference plots, using a best-fit algorithm.
5. The method for measuring an animal health of claim 1 wherein:
a frame rate for the series of image frames of step (a) is between 4 and 120 frames per second.
6. The method for measuring an animal health of claim 1 wherein:
the reduced vector field comprises between 24 and 600 vectors per frame.
7. The method for measuring an animal health of claim 1 wherein:
the first time period is between 15 seconds and 10 minutes.
8. The method for measuring an animal health of claim 1 wherein:
the second time period is the natural nocturnal activity period for the animal.
9. The method for measuring an animal health of claim 1 wherein:
the second time period is between 4 and 48 hours.
10. The method for measuring an animal health of claim 1 wherein:
the selecting a second-period maximum value set, step (f), comprises selecting n maximum values where n is between 4 and 100.
11. The method for measuring an animal health of claim 1 wherein:
the selecting a second-period maximum value set, step (f), comprises selecting maximum m % values where m % is between 1% and 15%.
12. A device for measuring animal health, comprising:
a camera, configured to record a series of image frames of an animal cage interior; and
a computer, configured to measure the animal health by performing steps (b)-(i) of the method of claim 1 .
13. The device for measuring animal health of claim 12 further comprising:
the animal cage;
wherein the animal cage contains the animal.
14. A method for measuring health of an animal in a medical study comprising the steps:
(a) recording, using a camera, a series of image frames of an animal cage interior wherein the cage comprises the animal enrolled in the medical study;
(b) generating a reduced vector field from at least a subset of image frames from step (a) using optical flow analysis;
(c) selecting one or more maximum vector values from the vector field from step (b);
(d) computing a first-period maximum value for the one or more maximum vector values from step (c) over a first time period length;
(e) repeating steps (a) through (d) for multiple, sequential time periods having the first time period length, wherein the multiple, sequential time periods fall within a second time period;
(f) selecting a second-period maximum value set from all of the first-period maximum values computed from repeating step (d);
(g) computing a mean value from the values of the second-period maximum value set selected from step (f);
(h) computing an expected time to die by subtracting a baseline value from the mean value from step (g); and
(i) plotting the computed expected time to die from step (h) over time.