CLASSIFICATION OF IMPACTS FROM SENSOR DATA

• US 20120150453A1
• Filed: 07/15/2011
• Published: 06/14/2012
• Est. Priority Date: 07/15/2010
• Status: Active Grant

• Current Assignee: Cleveland Clinic Foundation
• Original Assignee: Cleveland Clinic Foundation
• Inventors: Benzel, Edward C., Miele, Vincent J., Bartsch, Adam J.
• Granted Patent: US 9,289,176 B2
• Time in Patent Office: Days
• US Class Current: 702/41
• CPC Class Codes: A42B 3/046 Means for detecting hazards or accidents
• CPC Class Codes: A61B 2503/10 Athletes
• CPC Class Codes: A61B 2562/0219 Inertial sensors, e.g. accelerometers, gyroscopes, tilt switches
• CPC Class Codes: A61B 2562/0247 Pressure sensors
• CPC Class Codes: A61B 2562/0261 Strain gauges
• CPC Class Codes: A61B 2562/028 Microscale sensors, e.g. electromechanical sensors [MEMS]
• CPC Class Codes: A61B 5/1114 Tracking parts of the body
• CPC Class Codes: A61B 5/1121 Determining geometric values, e.g. centre of rotation or angular range of movement
• CPC Class Codes: A61B 5/1122 of movement trajectories
• CPC Class Codes: A61B 5/4064 Evaluating the brain A61B5/031, A61B5/369, A61B5/14553 take precedence
• CPC Class Codes: A61B 5/6803 Head-worn items, e.g. helmets, masks, headphones or goggles
• CPC Class Codes: A61B 5/682 Mouth, e.g., oral cavity; tongue; Lips; Teeth
• CPC Class Codes: A61B 5/7253 characterised by using transforms
• CPC Class Codes: A61B 5/7264 Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems neural networks per se G06N3/00; expert systems per se G06N5/00
• CPC Class Codes: A61B 5/7267 involving training the classification device
• CPC Class Codes: A61B 5/7278 Artificial waveform generation or derivation, e.g. synthesising signals from measured signals A61B5/327 takes precedence
• CPC Class Codes: A61B 5/7282 Event detection, e.g. detecting unique waveforms indicative of a medical condition A61B5/4094, A61B5/4818, A61B5/349, A61B5/0823 take precedence
• CPC Class Codes: A61B 5/746 Alarms related to a physiological condition, e.g. details of setting alarm thresholds or avoiding false alarms alarms per se G08B
• CPC Class Codes: G16H 50/20 for computer-aided diagnosis, e.g. based on medical expert systems

Abstract

Systems and methods are provided for classifying an impact to a head of a human being as one of a plurality of event classes. A sensor interface is configured to determine an acceleration at a location of interest on the human being from provided sensor data in response to the impact. A feature calculation component is configured to calculate a plurality of impact parameters from the determined acceleration at the location of interest. A pattern recognition classifier is configured to associate the impact with an associated event class of at least three associated event classes according to the calculated plurality of impact parameters. A post-processing component is configured to communicate the associated event class an observer via an associated output device.

Claims

1. A system for classifying an impact to a head of a human being as one of a plurality of event classes, the system comprising:

a sensor interface configured to determine an acceleration at a location of interest on the human being from provided sensor data in response to the impact;

a feature calculation component configured to calculate a plurality of impact parameters from the determined acceleration at the location of interest;

a pattern recognition classifier configured to associate the impact with an associated event class of at least three associated event classes according to the calculated plurality of impact parameters; and

a post-processing component configured to communicate the associated event class an observer via an associated output device.

2. The system of claim 1, further comprising a sensor assembly positioned at an associated measurement point on the head and configured to measure at least one of a linear acceleration, an angular acceleration, an angular velocity, and an orientation of the head at the associated measurement point during the impact to provide the sensor data, wherein the location of interest is remote from the associated measurement point.

3. The system of claim 2, wherein the sensor assembly is located in a mouth guard and a lip guard worn by the human being.

4. The system of claim 1, wherein the pattern recognition classifier is further configured to determine a desired range for values of the calculated impact parameters from historical data for the calculated impact parameters for the human being, the at least three event classes including a first class in which each of the calculated impact parameters are within their associated desired ranges, a second class in which none of the calculated impact parameters are within their associated desired ranges, and a third class in which at least one of the calculated impact parameters is within its associated desired range and at least one of the calculated impact parameters is not within its associated desired range.

5. The system of claim 1, wherein the at least three event classes represent ranges of likelihood for an injury to a structure within one of the head and neck given the measured impact.

6. The system of claim 1, wherein the at least three event classes represent respective mechanisms by which impacts are delivered, such that the associated event class represents a mechanism by which the impact was delivered.

7. The system of claim 1, wherein the location of interest is a center of gravity of the human being'"'"'s head.

8. The method of claim 1, wherein the location of interest is a location within one of a frontal lobe, parietal lobe, temporal lobe, occipital lobe, cerebellum, medulla oblongata, pons, thalamus, gyrus, formix, amygdyla, hippocampus, cranium, facial bones, maxilla, mandible, cerebrospinal fluid, occipito-cervical junction, cervical spine, vertebral bodies, spinal cord, spinal nerves, spinal vessels, basal ganglia, pen-vascular spaces, septum, white-gray matter junctions, bridging veins, corpus callosum, fissure, cavity sinus, meninges, falx, dura, arachnoid, and pia matter of the user.

9. The system of claim 1, wherein the pattern recognition classifier is one of an artificial neural network and a support vector machine.

10. The system of claim 1, wherein the calculated plurality of impact parameters includes the determined acceleration at a first time and the determined acceleration at a second time which is offset from the first time.

11. The system of claim 1, wherein the calculated plurality of impact parameters includes an angular acceleration at the location of interest.

12. The system of claim 1, wherein the calculated plurality of impact parameters includes a physical parameter calculated from a finite element model of the human being'"'"'s head and neck.

13. The system of claim 12, wherein the physical parameter is one of a Gadd Severity Index, a von Mises stress, and a Head injury Criterion, an impact force, a strain rate, a loading rate, a Generalized Acceleration Model for Brain Injury Threshold, a relative motion damage measure, a dilatational damage measure, a neck injury criterion, Head Impact Power parameter, a weighted principle component score, a strain energy, a cumulative strain damage measure, peak pressure, a principal strain, a sheer stress, and a product of a strain and the strain rate.

14. The system of claim 1, wherein the calculated plurality of impact parameters includes an angular velocity at the location of interest.

15. The system of claim 1, wherein the calculated plurality of impact parameters includes an angular position and orientation at the location of interest.

16. The system of claim 1, wherein the calculated plurality of impact parameters includes an angular jerk at the location of interest.

17. The system of claim 1, wherein the calculated plurality of impact parameters includes a linear jerk at the location of interest.

18. A method for determining a risk of injury to a human being comprising:

measuring at least one of a linear acceleration, an angular acceleration, an angular velocity and an orientation at a first location on the human being;

determining an acceleration at a location of interest on one of a head and a neck of the human being from the measured at least one of a linear acceleration, an angular acceleration, an angular velocity and an orientation at the first location, the first location being remote from the location of interest;

calculating a plurality of impact parameters from the determined acceleration at the location of interest;

associating the calculated plurality of impact parameters with an associated injury class of a plurality of injury classes, each injury class representing a range of probabilities that the human being will suffer an injury to a structure within one of the head and the neck given the calculated plurality of impact parameters; and

communicating the associated event class an observer via an associated output device.

19. The method of claim 18, wherein the plurality of injury classes represent ranges of probabilities of a concussion.

20. The method of claim 18, wherein the plurality of injury classes represent ranges of probabilities of a traumatic brain injury (TBI).

21. The method of claim 18, wherein the plurality of injury classes represent ranges of probabilities of a sub-concussion injury.

22. The method of claim 18, wherein the plurality of injury classes represent ranges of probabilities of a skull fracture.

23. The method of claim 18, wherein the plurality of injury classes represent ranges of probabilities of a facial fracture.

24. The method of claim 18, wherein the plurality of injury classes represent ranges of probabilities of a jaw fracture.

25. The method of claim 18, wherein the plurality of injury classes represent ranges of probabilities of a neck injury.

26. The method of claim 18, wherein the plurality of injury classes represent ranges of probabilities of a spinal cord injury.

27. The method of claim 18, wherein the plurality of injury classes include a first set of classes representing respective ranges of probabilities of a first injury type and a second set of classes representing respective ranges of probabilities of a second injury type.

28. A method for classifying an impact in a sporting event, the system comprising:

measuring at least one of a linear acceleration, an angular acceleration, an angular velocity and an orientation at a first location on an athlete participating in the sporting event;

determining an acceleration at a location of interest on one of a head and a neck of the athlete during the impact from the measured at least one of a linear acceleration, an angular acceleration, an angular velocity and an orientation at the first location, the first location being remote from the location of interest;

calculating a plurality of impact parameters from the determined acceleration at the location of interest;

associating the calculated plurality of impact parameters with an associated impact class of a plurality of impact classes, each impact class representing an associated type of impact within the sporting event; and

communicating the associated impact class an observer via an associated output device.

29. The method of claim 28, wherein the plurality of impact classes represent respective impact sources within the sporting event, such that the plurality of impact classes include at least one of a hook class, an oblique hook class, a jab class, an uppercut class, a cross class, and an overhand punch class.

30. The method of claim 28, wherein the plurality of impact classes represent a type and severity of an impact between two athletes, such that the plurality of impact classes include at least one of a helmet-to-helmet contact class, a helmet-to-knee class, a helmet-to-shoulder class, a helmet-to-foot class, a helmet-to-object class, a helmet-to-elbow class, a helmet-to-body class, and a helmet-to-ground class.

31. The method of claim 28, wherein the plurality of impact classes represent a type and severity of an impact between two players, such that the plurality of impact classes include a bare head-to-bare head contact class, a bare head-to-knee class, a bare head-to-elbow class, a bare head-to-foot class, a bare head-to-shoulder class, bare head-to-object class, bare head-to-body class and a bare head-to-ground class.

32. The method of claim 28, wherein the plurality of impact classes represent a handedness of a limb of an second athlete delivering the impact, such that the plurality of impact classes include a left hand class and a right hand class.

33. The method of claim 28, further comprising determining a desired range for values of the calculated impact parameters from historical data for the athlete, the plurality of impact classes comprising at least a first class representing impacts within the desired range and a second class representing impacts outside of the desired range.