Noise correcting patient fall risk state system and method for predicting patient falls

• US 9,318,012 B2
• Filed: 03/23/2012
• Issued: 04/19/2016
• Est. Priority Date: 12/12/2003
• Status: Active Grant

• Current Assignee: Careview Communications Incorporated
• Original Assignee: Kyle Brook Johnson, Stephen Ecker, Steve Gail Johnson
• Inventors: Johnson, Steve Gail, Johnson, Kyle Brook, Ecker, Stephen
• Primary Examiner(s): VO, TUNG T
• Application Number:
  US13/429,101
• Publication Number:
  US 20150199892A1
• Time in Patent Office: 1,488 Days
• Field of Search: 348/77
• US Class Current: 1/1
• CPC Class Codes: G08B 21/0476 Cameras to detect unsafe condition, e.g. video cameras
• CPC Class Codes: G08B 31/00 Predictive alarm systems characterised by extrapolation or other computation using updated historic data
• CPC Class Codes: H04N 7/183 for receiving images from a single remote source

Abstract

A patient fall prediction system from noise corrected surveillance video by identifying patient fall risk states. A hierarchy of discrete patient fall risk states, from no risk, to intermediate risk to critical risk, describe a patient fall risk. The system transitions from state to state based on changes detected in corresponding areas between a current video frame and a background frame. A set of fall risk state transition rules govern the entry into new fall risk states. A video frame is subdivided into multiple predetermined areas, at least two contain images of the patient. The number of false alarms are reduced by accurately defining fall risk state transition rules and by reducing the opportunity for noise to impact the state transition results. Frames that contain new changes are excluded from fall risk state processing, i.e., the first video frame that might cause an erroneous elevated fall risk state is culled.

Claims

1. A method for predicting a risk of a patient fall based on a plurality of fall risk states, the plurality of fall risk states defining a hierarchy of discrete fall risk states, at least one of the fall risk states being associated with an action, the method comprising:

receiving a plurality of video frames from a surveillance video camera, said surveillance video camera captures a sequence of video frames of a surveillance area, wherein a patient is present in the surveillance area, each of the plurality of video frames comprising a plurality of predetermined areas;

detecting a change in at least one of the plurality of predetermined areas of the current video frame of the plurality of video frames;

evaluating the change detected in the at least one of the plurality of predetermined areas of the current video frame for noise; and

identifying a fall risk transition to a fall risk state selected from the hierarchy of discrete fall risk states, the hierarchy of discrete fall risk states including a wait state, a no-wait state, and at least one other state, and wherein each fall risk state includes event and timing information, wherein the identification of a fall risk transition is based upon the change detected in the at least one of the plurality of predetermined areas of the current video frame.

2. The method recited in claim 1 above, wherein the plurality of fall risk states comprise a first fall risk state, and intermediate fall risk state wherein the patient is at a higher risk of a fall than the first fall risk state and a critical fall risk state, wherein the patient is at a higher risk of a fall than the first or intermediate fall risk states, wherein identifying a fall risk state from the hierarchy of discrete fall risk states from the change detected in the at least one of the plurality of predetermined areas of the current video frame further comprises:

entering the critical fall risk state from the intermediate fall risk state; and

issuing a fall alert based on entry into the critical fall risk state.

3. The method recited in claim 2 above, wherein identifying a fall risk state from the hierarchy of discrete fall risk states from the change detected in the at least one of the plurality of predetermined areas of the current video frame further comprises:

determining a current fall risk state of the hierarchy of discrete fall risk states, said current fall risk state being identified based on detecting a change in at least one of the plurality of predetermined areas of a background video frame of the plurality of video frames;

identifying a changed predetermined area of the plurality of predetermined areas;

accessing a set of fall risk state transition rules, said fall risk state transition rules define entry into a new fall risk state from a previous fall risk state based on an identity of a changed area;

identifying a fall state transition rule from the set of fall risk state transition rules for the current fall risk state and the identified changed predetermined area; and

entering a second fall risk state based on the identified fall state transition rule.

4. The method recited in claim 3 above, wherein detecting a change in at least one of the plurality of predetermined areas of the current video frame of the plurality of video frames, further comprises:

identifying a first plurality of pixels comprising a first predetermined area, a second plurality of pixels comprising a second predetermined area and a third plurality of pixels comprising a third predetermined area;

converting color information to a color value for each pixel of the first, second and third plurality of pixels of the first, second and third predetermined areas;

find a first current total area pixel weight of a color value all pixels for the first plurality of pixels, a second current total area pixel weight of a color value all pixels for the second plurality of pixels and a third current total area pixel weight of a color value all pixels for the third plurality in the respective first, second and third predetermined areas of the current video frame;

receiving a first background total area pixel weight of a color value all pixels in the first plurality of pixels, a second background total area pixel weight of a color value all pixels in the second plurality of pixels and a third background total area pixel weight of a color value all pixels in the third plurality of pixels of a corresponding one of the first, second and third predetermined areas of a background video frame; and

finding a first area difference total between the first current total area pixel weight and the first background total area pixel weight, a second area difference total between the second current total area pixel weight and the second background total area pixel weight and a third area difference total between the third current total area pixel weight and the third background total area pixel weight; and

comparing the first area difference total to a first threshold motion amount for the first predetermined area, second area difference total to a second threshold motion amount for the second predetermined area and third threshold motion amount for the third predetermined area of the current video frame.

5. The method recited in claim 1, wherein issuing a fall alert further comprises:

transmitting a remote location from the surveillance video camera; and

generating one of an audible alarm, a visual alarm, an audible message and a visual message at the remote location from the surveillance video camera.

6. The method recited in claim 1, wherein issuing a fall alert further comprises:

generating one of an audible alarm, a visual alarm, an audible message and a visual message locally to the surveillance video camera.

7. The method recited in claim 4, wherein the color information is one of chrominance and luminance.

8. The method recited in claim 4, wherein the color information is luminance.

9. The method recited in claim 4, wherein any predetermined area that exceeds the threshold motion amount is flagged as ACTIVE for the current video frame, the method further comprises:

storing an ACTIVE flag with any predetermined area that exceeds the threshold motion amount each.

10. The method recited in claim 9, wherein evaluating the change detected in the at least one of the plurality of predetermined areas of the current video frame for noise, further comprises:

receiving ACTIVE flags associated with predetermined areas of a background video frame;

identifying ACTIVE predetermined areas for the current video frame;

identifying an ACTIVE predetermined area for the current video frame without an ACTIVE flag for a corresponding predetermined area the background video frame; and

setting a fall risk state for the current video frame to the as the current fall risk state from the background video frame.

11. A system for predicting a risk of a patient fall based on a plurality of fall risk states, the plurality of fall risk states defining a hierarchy of discrete fall risk states, at least one of the fall risk states being associated with an action, the system comprising:

a surveillance video camera operative to capture a plurality of video frames of a surveillance area, wherein a patient is present in the surveillance area and each of the plurality of video frames comprises a plurality of predetermined areas; and

a video processor operative to detect a change in at least one of the plurality of predetermined areas of the current video frame of the plurality of video frames, evaluate the change detected in the at least one of the plurality of predetermined areas of the current video frame for noise, and identify a fall risk transition to a fall risk state selected from the hierarchy of discrete fall risk states, the hierarchy of discrete fall risk states including a wait state, a no-wait state, and at least one other state, and wherein each fall risk state includes event and timing information, wherein the identification of a fall risk transition is based upon the change detected in the at least one of the plurality of predetermined areas of the current video frame.

12. The system recited in claim 11 above, wherein the plurality of fall risk states comprise a first fall risk state, and intermediate fall risk state wherein the patient is at a higher risk of a fall than the first fall risk state and a critical fall risk state, wherein the patient is at a higher risk of a fall than the first or intermediate fall risk states, wherein identifying a fall risk state from the hierarchy of discrete fall risk states from the change detected in the at least one of the plurality of predetermined areas of the current video frame further comprises entering the critical fall risk state from the intermediate fall risk state and issuing a fall alert based on entry into the critical fall risk state.

13. The system recited in claim 12 above, wherein identifying a fall risk state from the hierarchy of discrete fall risk states from the change detected in the at least one of the plurality of predetermined areas of the current video frame further comprises:

determining a current fall risk state of the hierarchy of discrete fall risk states, said current fall risk state being identified based on detecting a change in at least one of the plurality of predetermined areas of a background video frame of the plurality of video frames;

identifying a changed predetermined area of the plurality of predetermined areas;

accessing a set of fall risk state transition rules, said fall risk state transition rules define entry into a new fall risk state from a previous fall risk state based on an identity of a changed area;

identifying a fall state transition rule from the set of fall risk state transition rules for the current fall risk state and the identified changed predetermined area; and

entering a second fall risk state based on the identified fall state transition rule.

14. The system recited in claim 13 above, wherein detecting a change in at least one of the plurality of predetermined areas of the current video frame of the plurality of video frames, further comprises:

identifying a first plurality of pixels comprising a first predetermined area, a second plurality of pixels comprising a second predetermined area and a third plurality of pixels comprising a third predetermined area;

converting color information to a color value for each pixel of the first, second and third plurality of pixels of the first, second and third predetermined areas;

find a first current total area pixel weight of a color value all pixels for the first plurality of pixels, a second current total area pixel weight of a color value all pixels for the second plurality of pixels and a third current total area pixel weight of a color value all pixels for the third plurality in the respective first, second and third predetermined areas of the current video frame;

receiving a first background total area pixel weight of a color value all pixels in the first plurality of pixels, a second background total area pixel weight of a color value all pixels in the second plurality of pixels and a third background total area pixel weight of a color value all pixels in the third plurality of pixels of a corresponding one of the first, second and third predetermined areas of a background video frame; and

finding a first area difference total between the first current total area pixel weight and the first background total area pixel weight, a second area difference total between the second current total area pixel weight and the second background total area pixel weight and a third area difference total between the third current total area pixel weight and the third background total area pixel weight; and

comparing the first area difference total to a first threshold motion amount for the first predetermined area, second area difference total to a second threshold motion amount for the second predetermined area and third threshold motion amount for the third predetermined area of the current video frame.

15. The system recited in claim 11, wherein issuing a fall alert further comprises:

transmitting a remote location from the surveillance video camera; and

generating one of an audible alarm, a visual alarm, an audible message and a visual message at the remote location from the surveillance video camera.

16. The system recited in claim 11, wherein issuing a fall alert further comprises generating one of an audible alarm, a visual alarm, an audible message and a visual message locally to the surveillance video camera.

17. The system recited in claim 14, wherein the color information is one of chrominance and luminance.

18. The system recited in claim 14, wherein the color information is luminance.

19. The system recited in claim 14, wherein any predetermined area that exceeds the threshold motion amount is flagged as ACTIVE for the current video frame, the method further comprises:

storing an ACTIVE flag with any predetermined area that exceeds the threshold motion amount each.

20. The system recited in claim 19, wherein evaluating the change detected in the at least one of the plurality of predetermined areas of the current video frame for noise, further comprises:

receiving ACTIVE flags associated with predetermined areas of a background video frame;

identifying ACTIVE predetermined areas for the current video frame;

identifying an ACTIVE predetermined area for the current video frame without an ACTIVE flag for a corresponding predetermined area the background video frame; and

setting a fall risk state for the current video frame to the as the current fall risk state from the background video frame.