Kalman filter approach to augment object tracking

1. A computer-implemented method for tracking one or more objects at each position in an interval in a video input with the use of a Kalman filter having a noisy measurement component and a noisy predictor component, comprising:

obtaining a first location estimate of an object in the video input at each position in the interval, wherein the first location estimate represents the noisy measurement component of the Kalman filter;

obtaining a second location estimate and a movement estimate of the object in the video input at each position in the interval, wherein the second location estimate and movement estimate represent the noisy predictor component of the Kalman filter; and

determining a final estimate of at least one selected from the group consisting of (1) a location of the object at a future interval at each position in the interval with the Kalman filter and (2) a velocity of the object at a future interval at each position in the interval with the Kalman filter,

wherein each final estimate is determined with reference to a Kalman gain, the Kalman gain indicating a first level of importance of the noisy measurement component and a second level of importance of the noisy predictor component, and

wherein the first location estimate, the second location estimate, and the movement estimate each comprise estimates having two dimensional components in a Cartesian coordinate system, the two dimensional components consisting of an x dimension and a y dimension.

2. The method of claim 1 , wherein the Kalman filter uses a steady-state Kalman filter and uses a predetermined Kalman gain.

3. The method of claim 2 , further comprising:

obtaining a first location estimate of an object in at least one separate video input at each position in an interval and determining a plurality of errors of the first location estimate;

obtaining a second location estimate and a movement estimate of the object in the at least one separate video input at each position in the interval and determining a plurality of errors of the second location estimate and movement estimate;

estimating the predetermined Kalman gain according to at least one variance of the plurality of errors of the first location estimate and at least one variance of the plurality of errors of the second location estimate and movement estimate.

4. The method of claim 1 , wherein the first location estimate, the second location estimate, and the movement estimate each comprise estimates having a single dimensional component.

5. The method of claim 1 , wherein the first location estimate, the second location estimate, and the movement estimate each comprise estimates having a plurality dimensional components.

6. The method of claim 1 , wherein the x dimension and the y dimension of the first location estimate, the second location estimate, and the movement estimate are determined independently.

7. The method of claim 1 , wherein the first location estimate represents a measurement x t O , where t is a position in the interval, wherein the second position estimate and movement estimate are represented by a prediction process {tilde over (x)} t , and wherein determining the final estimate further comprises:

providing a predetermined Kalman gain K (K 1 , K 2 ), where K 1 represents the first level of importance of the noisy measurement component, and where K 2 represents the second level of importance of the noisy predictor component;

determining an estimate of a position x t and an estimate of velocity {dot over (x)} t , at each position t, where each component of the position is measured independently for an x coordinate and a y coordinate of the Cartesian space,

wherein the position x, and the velocity is determined according to the constraints

{circumflex over (x)} t =K 1 x t O +(1− K 1 ) {tilde over (x)} t and {dot over ({circumflex over ( x )} t ={dot over ({tilde over ( x )} t +K 2 ( x t O −{tilde over (x)} t ).

8. The method of claim 1 , wherein the object comprises a human face.

9. A computer program product stored on a non-transitory computer readable storage device for tracking one or more objects at each position in an interval in a video input with the use of a Kalman filter having a noisy measurement component and a noisy predictor component, which when executed, causes a computer to perform a method, comprising:

obtaining a first location estimate of an object in the video input at each position in the interval, wherein the first location estimate represents the noisy measurement component of the Kalman filter;

obtaining a second location estimate and a movement estimate of the object in the video input at each position in the interval, wherein the second location estimate and movement estimate represent the noisy predictor component of the Kalman filter; and

determining a final estimate of at least one selected from the group consisting of (1) a location of the object at a future interval at each position in the interval with the Kalman filter and (2) a velocity of the object at a future interval at each position in the interval with the Kalman filter,

wherein each final estimate is determined with reference to a Kalman gain, the Kalman gain indicating a first level of importance of the noisy measurement component and a second level of importance of the noisy predictor component, and

wherein the first location estimate, the second location estimate, and the movement estimate each comprise estimates having two dimensional components in a Cartesian coordinate system, the two dimensional components consisting of an x dimension and a y dimension.

10. The computer program product of claim 9 , wherein the method further comprises:

providing a display device;

displaying on the display device a rendering of the video input and a plurality of information of the final estimate of position and/or velocity.

11. The computer program product of claim 9 , wherein the rendering of the video input comprises a visual array of a plurality of frames of the input video, each of the plurality of frames corresponding to one position in the interval.

12. The computer product of claim 9 , wherein the plurality of information is superimposed on the rendering of the video input.

13. An object tracking system for tracking one or more objects at each position in an interval in a video input with the use of a Kalman filter having a noisy measurement component and a noisy predictor component, comprising:

at least one memory for storing the video input;

at least one processor in communication with the at least one memory;

an object detecting system comprising executable code stored in the at least one memory configured to instruct the at least one processor to obtain a first location estimate of an object in the video input at each position in the interval, wherein the first location estimate represents the noisy measurement component in the Kalman filter,

and to obtain a second location estimate and a movement estimate of the object in the video input at each position in the interval, wherein the second location estimate and movement estimate represent the noisy predictor component in the Kalman filter; and

a final estimate system comprising executable code stored in to the memory configured to instruct the at least one processor to determine a final estimate of a position and/or a velocity of the object at a future time at each position in the interval with the Kalman filter, wherein each final estimate is determined with reference to a Kalman gain, the Kalman gain indicating a first level of importance of the noisy measurement component and a second level of importance of the noisy predictor component, and wherein the first location estimate, the second location estimate, and the movement estimate each comprise estimates having two dimensional components in a Cartesian coordinate system, the two dimensional components consisting of an x dimension and a y dimension.

14. The system of claim 13 , further comprising:

a video capture device operatively connected to the at least one memory, whereby the video capture device captures the video input and stores the video input in the at least one memory.

15. The system of claim 13 , further comprising:

a display device operatively connected to the at least one processor, adapted to display a rendering of the video input and a plurality of information of the final estimate of position and/or velocity.

16. The method of claim 13 , wherein the video input is generated from a video input device.

17. The method of claim 13 , wherein the video input device is a security camera.