Multicamera Calibration via Human Pose and Cross-Ratio Invariants
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current multicamera calibration techniques are burdensome and require preprinted patterns or tags, making them inaccessible to non-technical users and limiting the portability and accessibility of multicamera applications, especially when determining extrinsic parameters such as 3D positions and orientations of cameras.
Innovation Solution
The use of a human subject as a calibration pattern, employing cross-ratio invariants to calculate transformations between cameras without the need for preprinted patterns, allowing users to calibrate multicamera systems using anatomical points and machine learning models to identify and estimate image coordinates, enabling extrinsic parameter determination without knowing the intrinsic camera parameters.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If preprinted patterns or tags are used for calibration, then calibration accuracy can be achieved, but the process becomes burdensome and inaccessible to non-technical users
Solution Approach 1:
The patent replaces expensive, specialized calibration patterns with a human subject that is universally available. The human body serves as a temporary, disposable calibration object that eliminates the need for costly preprinted patterns or tags, making calibration accessible to non-technical users while maintaining accuracy through anatomical landmark detection
Solution Approach 2:
The calibration system uses the human subject's own anatomical features (landmarks, joints, body proportions) as the calibration reference, eliminating the need for external calibration objects. The system automatically detects and processes these self-contained features through machine learning models, enabling users to calibrate without specialized knowledge or equipment
2Measurement precision
If preprinted patterns are used for calibration, then extrinsic parameters can be determined, but the calibration process becomes complex and requires specialized equipment
Solution Approach 1:
The patent replaces mechanical calibration systems (physical patterns, tags, and measurement devices) with a computational approach using machine learning models. The system processes images of human anatomical landmarks and automatically calculates extrinsic parameters through algorithmic analysis, eliminating the need for physical calibration equipment and reducing operational complexity
Solution Approach 2:
The patent changes the fundamental calibration parameter from external patterns to internal human anatomical features. By using body proportions, landmark positions, and skeletal relationships as calibration references, the system transforms the calibration process from equipment-dependent to biology-dependent, simplifying the required apparatus while maintaining measurement precision
3Area of stationary object
If small calibration patterns are used, then calibration can be performed in limited space, but the calibration distance is restricted
Solution Approach 1:
The patent employs a dynamic calibration object (human subject) that can adapt its pose and position to suit different calibration scenarios. The human can assume various configurations (T-pose, A-pose, dynamic movements) that provide sufficient geometric diversity for calibration at varying distances, eliminating the fixed-size limitation of traditional patterns while maintaining space efficiency
Data Source
Figure 1A~1B
Figure 2
Figure 3
AI summary
Methods and apparatus to calibrate a multicamera system based on a human pose are disclosed. An example apparatus includes an object identifier to identify a first set of coordinates defining first locations of anatomical points of a human in a first image captured by a first camera and identify a second set of coordinates defining second locations of the anatomical points in a second image captured by a second camera. The apparatus includes a pose detector to detect, based on at least one of the first or second sets of coordinates, when the human is in a particular pose. The apparatus includes a transformation calculator to, in response to detection of the human in the particular pose, calculate a relative transformation between the first camera and the second camera based on a first subset of the first set of coordinates and a second subset of the second set of coordinates.