System and method for the autonomous identification of physical abuse
Embodiments of the present systems and methods may provide techniques that may provide the capability to autonomously and automatically identify cases of suspected abuse, which may then be subject to follow up review and action. For example, in an embodiment, a system for identifying potential physical abuse in a location may comprise at least one video camera provided at the location and a computer system comprising a processor, memory accessible by the processor, and computer program instructions stored in the memory and executable by the processor to cause the computer system to perform: receiving video data from the at least one video camera, identifying potential physical abuse from the video data, the identification being to within a predetermined confidence level, and transmitting information indicating that a potential human threat has been identified.
1. A system for identifying potential physical abuse in a location comprising:
at least one video camera provided at the location; and
a computer system comprising a processor, a memory accessible by the processor, and computer program instructions stored in the memory and executable by the processor to cause the computer system to perform:
receiving video data from the at least one video camera,
identifying potential physical abuse from the video data, the identification of potential physical abuse being to within a predetermined confidence level, and
transmitting information indicating that a potential human threat has been identified
wherein identifying potential physical abuse comprises:
tracking and modeling bodily motion of at least one person shown in the video data, including identifying and tracking 2D joint center positions of the at least one person;
quantifying velocity and acceleration of at least one portion of a body of the at least one person shown in the video data using the tracked and modeled bodily motion; and
determining potential physical abuse based on a comparison of the quantified velocity and acceleration of the at least one portion of the body of the at least one person shown in the video data with at least one threshold.
2. The system of claim 1 , wherein the tracking and modeling bodily motion of the at least one person shown in the video data further comprises:
predicting a motion of the at least one portion of the body of the at least one person shown in the video data on a 3D human model as joint angles of the at least one person over time.
3. The system of claim 1 , wherein the predicting the motion of at the least one portion of the body comprises:
connecting 2D joint center positions of the at least one person shown in the video data to form line segments.
4. The system of claim 1 , wherein the at least one portion of the body comprises at least one of a wrist, a shoulder, an elbow, a foot, or a leg.
5. A method for identifying potential physical abuse, implemented in a computer comprising a processor, a memory accessible by the processor, and computer program instructions stored in the memory and executable by the processor, the method comprising:
receiving video data from at least one video camera at a location;
identifying potential physical abuse from the video data, the identification of potential physical abuse being to within a predetermined confidence level; and
transmitting information indicating that a potential human threat has been identified,
wherein identifying potential physical abuse comprises:
tracking and modeling bodily motion of at least one person shown in the video data, including identifying and tracking 2D joint center positions of the at least one person;
quantifying velocity and acceleration of at least one portion of a body of the at least one person shown in the video data using the tracked and modeled bodily motion; and
determining potential physical abuse based on a comparison of the quantified velocity and acceleration of the at least one portion of the body of the at least one person shown in the video data with at least one threshold.
6. The method of claim 5 , wherein the tracking and modeling bodily motion of the at least one person shown in the video data further comprises:
predicting motion of the at least one portion of the body of the at least one person shown in the video data on a 3D human model as joint angles of the at least one person over time.
7. The method of claim 5 , wherein the predicting the motion of the at least one portion of the body comprises:
connecting 2D joint center positions of the at least one person shown in the video data to form line segments.
8. The method of claim 5 , wherein the at least one portion of the body comprises at least one of a wrist, a shoulder, an elbow, a foot, or a leg.