Video-based fall risk assessment system

Abstract

Various embodiments of a video-based fall risk assessment system are disclosed. During operation, this fall risk assessment system can receives a sequence of video frames including a person being monitored for fall risk assessment. The system next generates a sequence of action labels for the sequence of video frames by, for each video frame in the sequence of video frames: estimating a pose of the person within the video frame; and classifying the estimated pose as a given action among a set of predetermined actions. Next, the system identifies a subset of action labels within the sequence of action labels. The system next extracts a set of gait features for the person from a subset of video frames within the sequence of video frames corresponding to the subset of action labels. Subsequently, the system analyzes the set of extracted gait features to generate a fall risk assessment for the person. In some embodiments, the sequence of video frames is captured during a predetermined time period, such as an hour, a day, or a week.

What is claimed is: 
     
         1 . A method of performing video-based fall risk assessment, comprising:
 receiving a sequence of video frames including a person being monitored for fall risk assessment;   generating a sequence of action labels for the sequence of video frames by, for each video frame in the sequence of video frames:
 estimating a pose of the person within the video frame; and 
 classifying the estimated pose as a given action among a set of predetermined actions; 
   identifying a subset of action labels within the sequence of action labels;   extracting a set of gait features for the person from a subset of video frames within the sequence of video frames corresponding to the subset of action labels; and   analyzing the set of extracted gait features to generate a fall risk assessment for the person.   
     
     
         2 . The method of  claim 1 , wherein the sequence of video frames is captured during a predetermined time period. 
     
     
         3 . The method of  claim 2 , wherein the predetermined time period is an hour, a day, or a week. 
     
     
         4 . The method of  claim 1 , wherein prior to estimating a pose of the person within the video frame, the method further comprises detecting the person within the video frame. 
     
     
         5 . The method of  claim 1 , wherein the set of predetermined actions a standing action, a sitting action, a walking action, and all other actions. 
     
     
         6 . The method of  claim 5 , wherein identifying the subset of action labels within the sequence of action labels includes identifying all action labels classified the walking action. 
     
     
         7 . The method of  claim 1 , wherein the set of gait features includes one or more of: step count, average step duration, variance of step duration for one foot or both feet, speed, cadence, step balance, and body sway factor. 
     
     
         8 . The method of  claim 2 , wherein analyzing the set of extracted gait features to generate a fall risk assessment for the person includes analyzing the sequence of video frames captured during the predetermined time period. 
     
     
         9 . The method of  claim 1 , wherein analyzing the set of extracted gait features to generate a fall risk assessment includes perform one or more statistical analyses on a given extracted gait feature in the set of extracted gait features. 
     
     
         10 . The method of  claim 1 , wherein the method further comprises triggering a high-fall-risk warning to be sent to the caregivers when analyzing the set of extracted gait features generates a high-fall-risk assessment for the person. 
     
     
         11 . A video-based fall risk assessment system, comprising:
 one or more processors;   a memory coupled to the one or more processors,   wherein the memory stores instructions that, when executed by the one or more processors, cause the system to:
 receive a sequence of video frames including a person being monitored for fall risk assessment; 
 generate a sequence of action labels for the sequence of video frames by, for each video frame in the sequence of video frames:
 estimating a pose of the person within the video frame; and 
 classifying the estimated pose as a given action among a set of predetermined actions; 
 
 identify a subset of action labels within the sequence of action labels; 
 extract a set of gait features for the person from a subset of video frames within the sequence of video frames corresponding to the subset of action labels; and 
 analyze the set of extracted gait features to generate a fall risk assessment for the person. 
   
     
     
         12 . The system of  claim 11 , wherein the sequence of video frames is captured during a predetermined time period. 
     
     
         13 . The system of  claim 12 , wherein the predetermined time period is an hour, a day, or a week. 
     
     
         14 . The system of  claim 11 , wherein the memory further stores instructions that, when executed by the one or more processors, cause the system to detect the person within the video frame prior to estimating a pose of the person within the video frame. 
     
     
         15 . The system of  claim 11 , wherein the set of predetermined actions a standing action, a sitting action, a walking action, and all other actions. 
     
     
         16 . The system of  claim 15 , wherein identifying the subset of action labels within the sequence of action labels includes identifying all action labels classified the walking action. 
     
     
         17 . The system of  claim 11 , wherein the set of gait features includes one or more of: step count, average step duration, variance of step duration for one foot or both feet, speed, cadence, step balance, and body sway factor. 
     
     
         18 . The system of  claim 12 , wherein analyzing the set of extracted gait features to generate a fall risk assessment includes perform one or more statistical analyses on a given extracted gait feature in the set of extracted gait features. 
     
     
         19 . The system of  claim 11 , wherein the memory further stores instructions that, when executed by the one or more processors, cause the system to trigger a high-fall-risk warning to be sent to the caregivers when analyzing the set of extracted gait features generates a high-fall-risk assessment for the person. 
     
     
         20 . An embedded system, comprising:
 one or more cameras configured to capture a sequence of video frames including a person;   one or more processors;   a memory coupled to the one or more processors and storing instructions that, when executed by the one or more processors, cause the system to:
 receive a sequence of video frames including a person being monitored for fall risk assessment; 
 generate a sequence of action labels for the sequence of video frames by, for each video frame in the sequence of video frames:
 estimating a pose of the person within the video frame; and 
 classifying the estimated pose as a given action among a set of predetermined actions; 
 
 identify a subset of action labels within the sequence of action labels; 
 extract a set of gait features for the person from a subset of video frames within the sequence of video frames corresponding to the subset of action labels; and 
 analyze the set of extracted gait features to generate a fall risk assessment for the person.