Body Posture Tracking

• US 20140358475A1
• Filed: 05/27/2014
• Published: 12/04/2014
• Est. Priority Date: 05/29/2013
• Status: Active Grant

• Current Assignee: Dassault Systemes SA
• Original Assignee: Dassault Systemes SA
• Inventors: Boulkenafed, Malika, Michel, Fabrice Francis
• Granted Patent: US 10,830,584 B2
• Time in Patent Office: Days
• US Class Current: 702/152
• CPC Class Codes: G01B 21/16 for measuring distance of clearance between spaced objects
• CPC Class Codes: G06F 18/256 of results relating to different input data, e.g. multimodal recognition
• CPC Class Codes: G06F 3/017 Gesture based interaction, e.g. based on a set of recognized hand gestures interaction based on gestures traced on a digitiser G06F3/04883
• CPC Class Codes: G06V 10/811 the classifiers operating on different input data, e.g. multi-modal recognition
• CPC Class Codes: G06V 40/23 Recognition of whole body movements, e.g. for sport training

Abstract

It is provided a computer-implemented method for performing body posture tracking, comprising the steps of collecting (S10) depth measurements of a body with at least one depth sensor; collecting (S20) inertial measurements with at least one inertial sensor attached to the body; and determining (S30) at least one posture of the body as a function of the depth measurements and the inertial measurements. Such a method improves the field of body posture tracking.

Claims

1. A computer-implemented method for performing body posture tracking, comprising the steps of:

collecting (S10) depth measurements (Z_t+1) of a body with at least one depth sensor;

collecting (S20) inertial measurements (X_t+1) with at least one inertial sensor attached to the body; and

determining (S30) at least one posture (y*) of the body as a function of the depth measurements and the inertial measurements.

2. The method of claim 1, wherein the depth measurements and the inertial measurements constitute time-series, and determining a posture is repeated at each time the method collects depth measurements and/or inertial measurements, the method thereby tracking body motion.

3. The method of claim 2, wherein determining a posture at each current time (t+1) is performed based on predetermined data, on the depth measurements (z_t+1) and/or inertial measurements (x_t+1) collected at the current time, and on data determined at the previous time.

4. The method of claim 3, wherein determining the posture comprises determining and maximizing a probability distribution (P(y_t+1|X_t+1, Z_t+1)) that assigns probabilities to postures (y) of the body as a function of the depth measurements and the inertial measurements.

5. The method of claim 4 wherein, at each current time (t+1), the probability distribution (P(y_t+1|X_t+1, Z_t+1)) that assigns probabilities to postures (y) of the body as a function of the depth measurements and the inertial measurements is recurrently obtained by multiplying the probability distribution (P(y_t|X_t, Z_t)) that assigns probabilities to postures (y) of the body as a function of the depth measurements and the inertial measurements determined at the previous time (t), by a probability distribution (P_inertial(y_t+1|x_t+1)) that assigns probabilities to postures (y) of the body as a function of the inertial measurements only and/or a probability distribution (P_depth(y_t+1|z_t+1)) that assigns probabilities to postures (y) of the body as a function of the depth measurements only, the probability distribution function of the inertial measurements only and/or the depth measurements only being obtained based on the predetermined data.

6. The method of claim 1, wherein the method comprises an initial stage, including the steps of:

tracking (S150) true states (y) of a body taking postures;

collecting (S100) depth measurements (z) of the body with said at least one depth sensor, and collecting (S200) inertial measurements (x) with said at least one inertial sensor attached to the body; and

then establishing (S300) the function that provides said at least one posture (y*) of the body for any values of depth measurements (z) and inertial measurements (x).

7. The method of claim 6, wherein the depth measurements and the inertial measurements constitute time-series and determining a posture is repeated at each time the method collects depth measurements and/or inertial measurements, and establishing the function comprises determining data based on which determining a posture at each current time (t+1) is performed.

8. The method of claim 7, wherein the determined data comprise data based on which the method determines a probability distribution (P_inertial(y_t+1|X_t+1)) that assigns probabilities to postures (y) of the body as a function of the inertial measurements only and a probability distribution (P_depth(y_t+1|z_t+1)) that assigns probabilities to postures (y) of the body as a function of the depth measurements only.

9. A computer-implemented method constituting an initial stage of a method for performing body posture tracking, comprising the steps of:

tracking true states (y) of a body taking postures;

collecting depth measurements (z) of the body with at least one depth sensor, and collecting inertial measurements (x) with at least one inertial sensor attached to the body; and

then

establishing a function that provides at least one posture (y*) of the body for any values of depth measurements (z) and inertial measurements (x).

10. The method of claim 9, wherein establishing the function comprises determining data based on which, if for the method for performing body posture tracking, the depth measurements and the inertial measurements constitute time-series and determining a posture is repeated at each time the method collects depth measurements and/or inertial measurements, determining a posture at each current time (t+1) may be performed.

11. The method of claim 10, wherein the determined data comprise data based on which the method for performing body posture tracking may determine a probability distribution (P_inertial(y_t+1|x_t+1)) that assigns probabilities to postures (y) of the body as a function of the inertial measurements only and/or a probability distribution (P_depth(y_t+1|z_t+1)) that assigns probabilities to postures (y) of the body as a function of the depth measurements only.

12. A computer program product comprising:

a data storage medium embodying code instructions for performing body posture tracking; and

the code instructions including instructions configuring a computer to;

collect (S10) depth measurements (Z_t+1) of a body with at least one depth sensor;

collect (S20) inertial measurements (X_t+1) with at least one inertial sensor attached to the body; and

determine (S30) at least one posture (y*) of the body as a function of the depth measurements and the inertial measurements.

13. A system comprising:

a processor; and

a memory operatively coupled to the processor and having recorded thereon program code, wherein the processor executing the program code;

collects (S10) depth measurements (Z_t+1) of a body with at least one depth sensor;

collects (S20) inertial measurements (X_t+1) with at least one inertial sensor attached to the body; and

determines (S30) at least one posture (y*) of the body as a function of the depth measurements and the inertial measurements.

14. The system of claim 13, further comprising at least one depth sensor and at least one inertial sensor.

15. The system of claim 13, wherein the depth measurements and the inertial measurements constitute time-series, and the processor determines a posture by repeating at each time that the processor collects depth measurements and/or inertial measurements, the system thereby tracking body motion.

16. The system of claim 15, wherein the processor determines a posture at each current time (t+1) based on predetermined data, on the depth measurements (z_t+1) and/or inertial measurements (x_t+1) collected at the current time, and on data determined at the previous time.

17. The system of claim 16, wherein the processor determines the posture by:

determining and maximizing a probability distribution (P(y_t+1|X_t+1, Z_t+1)) that assigns probabilities to postures (y) of the body as a function of the depth measurements and the inertial measurements.

18. The system of claim 13, wherein the processor executing the program code further performs an initial stage to:

track (S150) true states (y) of a body taking postures;

collect (S100) depth measurements (z) of the body with said at least one depth sensor, and collect (S200) inertial measurements (x) with said at least one inertial sensor attached to the body; and

then

establish (S300) the function that provides said at least one posture (y*) of the body for any values of depth measurements (z) and inertial measurements (x).

19. The system of claim 18, wherein the depth measurements and the inertial measurements constitute time-series and determining a posture is repeated at each time the method collects depth measurements and/or inertial measurements, and establishing the function comprises determining data based on which determining a posture at each current time (t+1) is performed.

20. The system of claim 19, wherein the determined data comprise data based on which the method determines a probability distribution (P_inertial(y_t+1|x_t+1)) that assigns probabilities to postures (y) of the body as a function of the inertial measurements only and a probability distribution (P_depth(y_t+1|z_t+1)) that assigns probabilities to postures (y) of the body as a function of the depth measurements only.