Technology to automatically identify the frontal body orientation of individuals in real-time multi-camera video feeds

Patent metadata

Field	Value
Publication number	US-12118775-B2
Application number	US-201917438098-A
Country	US
Kind code	B2
Filing date	Apr 12, 2019
Priority date	Apr 12, 2019
Publication date	Oct 15, 2024
Grant date	Oct 15, 2024

Abstract

Methods, systems and apparatuses may provide for technology that detects an individual in a real-time multi-camera video feed and generates three-dimensional (3D) skeletal data based on the real-time multi-camera video feed. The technology may also automatically identify a frontal body orientation of an individual based on the 3D skeletal data and one or more anthropometric constraints.

First claim

We claim:

1. A performance-enhanced computing system comprising:
a network controller to receive a real-time multi-camera video feed;
a processor including logic coupled to one or more substrates, wherein the logic is implemented at least partly in one or more of configurable logic or fixed-functionality hardware logic, the logic coupled to the one or more substrates to:
detect an individual in the real-time multi-camera video feed,
generate three-dimensional (3D) skeletal data that includes shoulder coordinates based on the real-time multi-camera video feed, wherein the 3D skeletal data is to correspond to the individual, and
automatically identify a frontal body orientation of the individual based on the 3D skeletal data and one or more anthropometric constraints by determining a shoulder-to-shoulder line of the individual based on a left shoulder coordinate and a right shoulder coordinate, and determining a perspective line relative to the shoulder-to-shoulder line that is in a direction of the frontal body orientation; and

an immersive video subsystem to generate a video frame from a perspective of the individual based on the frontal body orientation and the real-time multi-camera video feed.

2. The computing system of claim 1 , wherein the 3D skeletal data further includes a neck base coordinate and a head coordinate, wherein to determine the perspective line, the logic coupled to the one or more substrates is to determine a neck tilt of the individual based on the head coordinate relative to the neck base coordinate, and wherein the one or more anthropometric constraints indicate that the frontal body orientation is toward the neck tilt.

3. The computing system of claim 1 , wherein the 3D skeletal data further includes a left knee coordinate, a left hip coordinate, a right knee coordinate, and a right hip coordinate, and wherein to determine the perspective line, the logic coupled to the one or more substrates is to:
determine a left knee bend direction of the individual based on the left knee coordinate relative to the left hip coordinate, and
determine a right knee bend direction of the individual based on the right knee coordinate relative to the right hip coordinate, wherein the one or more anthropometric constraints indicate that the frontal body orientation is toward one or more of the left knee bend direction or the right knee bend direction.

4. The computing system of claim 1 , wherein the 3D skeletal data further includes a left elbow coordinate and a right elbow coordinate, wherein to determine the perspective line, the logic coupled to the one or more substrates is to:
determine a left elbow rotation of the individual based on the left elbow coordinate relative to the left shoulder coordinate, and
determine a right elbow rotation of the individual based on the right elbow coordinate relative to the right shoulder coordinate, wherein the one or more anthropometric constraints indicate that the frontal body orientation is away from one or more of the left elbow rotation or the right elbow rotation.

5. The computing system of claim 1 , wherein the logic coupled to the one or more substrates is to reconcile one or more conflicting limb parts in the 3D skeletal data.

6. A semiconductor apparatus comprising:
one or more substrates; and
logic coupled to the one or more substrates, wherein the logic is implemented at least partly in one or more of configurable logic or fixed-functionality hardware logic, the logic coupled to the one or more substrates to:
detect an individual in a real-time multi-camera video feed;
generate three-dimensional (3D) skeletal data that includes shoulder coordinates based on the real-time multi-camera video feed, wherein the 3D skeletal data is to correspond to the individual; and

automatically identify a frontal body orientation of the individual based on the 3D skeletal data and one or more anthropometric constraints by determining a shoulder-to-shoulder line of the individual based on a left shoulder coordinate and a right shoulder coordinate, and determining a perspective line relative to the shoulder-to-shoulder line that is in a direction of the frontal body orientation.

7. The semiconductor apparatus of claim 6 , wherein the 3D skeletal data further includes a neck base coordinate and a head coordinate, wherein to determine the perspective line, the logic coupled to the one or more substrates is to determine a neck tilt of the individual based on the head coordinate relative to the neck base coordinate, and wherein the one or more anthropometric constraints indicate that the frontal body orientation is toward the neck tilt.

8. The semiconductor apparatus of claim 6 , wherein the 3D skeletal data further includes a left knee coordinate, a left hip coordinate, a right knee coordinate, and a right hip coordinate, and wherein to determine the perspective line, the logic coupled to the one or more substrates is to:
determine a left knee bend direction of the individual based on the left knee coordinate relative to the left hip coordinate; and
determine a right knee bend direction of the individual based on the right knee coordinate relative to the right hip coordinate, wherein the one or more anthropometric constraints indicate that the frontal body orientation is toward one or more of the left knee bend direction or the right knee bend direction.

9. The semiconductor apparatus of claim 6 , wherein the 3D skeletal data further includes a left elbow coordinate and a right elbow coordinate, wherein to determine the perspective line, the logic coupled to the one or more substrates is to:
determine a left elbow rotation of the individual based on the left elbow coordinate relative to the left shoulder coordinate; and
determine a right elbow rotation of the individual based on the right elbow coordinate relative to the right shoulder coordinate, wherein the one or more anthropometric constraints indicate that the frontal body orientation is away from one or more of the left elbow rotation or the right elbow rotation.

10. The semiconductor apparatus of claim 6 , wherein the logic coupled to the one or more substrates is to reconcile one or more conflicting limb parts in the 3D skeletal data.

11. At least one computer readable storage medium comprising a set of instructions, which when executed by a computing system, cause the computing system to:
detect an individual in a real-time multi-camera video feed;
generate three-dimensional (3D) skeletal data that includes shoulder coordinates based on the real-time multi-camera video feed, wherein the 3D skeletal data is to correspond to the individual; and
automatically identify a frontal body orientation of the individual based on the 3D skeletal data and one or more anthropometric constraints by determining a shoulder-to-shoulder line of the individual based on a left shoulder coordinate and a right shoulder coordinate, and determining a perspective line relative to the shoulder-to-shoulder line that is in a direction of the frontal body orientation.

12. The at least one computer readable storage medium of claim 11 , wherein the 3D skeletal data further includes a neck base coordinate and a head coordinate, wherein to determine the perspective line the instructions, when executed, cause the computing system to determine a neck tilt of the individual based on the head coordinate relative to the neck base coordinate, and wherein the one or more anthropometric constraints indicate that the frontal body orientation is toward the neck tilt.

13. The at least one computer readable storage medium of claim 11 , wherein the 3D skeletal data further includes a left kn

Assignees

• Intel Corp

Inventors

• Li Qiang
• Li Wenlong
• Wang Yumeng
• Tong Xiaofeng

Classifications

• G06V40/20

  Movements or behaviour, e.g. gesture recognition (recognition of facial expressions G06V40/16) · CPC title

• G06T17/10

  Constructive solid geometry [CSG] using solid primitives, e.g. cylinders, cubes · CPC title

• G06V40/23

  Recognition of whole body movements, e.g. for sport training · CPC title

• G06V40/103

  Static body considered as a whole, e.g. static pedestrian or occupant recognition · CPC title

• G06V20/52

  Surveillance or monitoring of activities, e.g. for recognising suspicious objects (recognising microscopic objects G06V20/69) · CPC title