Contactless Degree of Freedom Determination Using Rad-D Image Fusion

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current methods for determining the internal degrees of freedom of flexible bodies in three-dimensional space, such as human pose estimation, face limitations in quality and flexibility due to the ambiguity of 3D coordinate reconstruction from radiometric and depth images, requiring complex models and high computational effort, and are often limited to specific object classes and lighting conditions.

Innovation Solution

A method that combines radiometric and depth images to determine both external and internal degrees of freedom by creating a rad-D image, using a transformation to assign depth values to radiometric image pixels, identifying characteristic object points, and calculating spatial coordinates, allowing for the establishment of a body coordinate system and deformation transformations to accurately represent the spatial state of flexible bodies.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If 3D coordinate reconstruction is performed from radiometric and depth images, then spatial state determination is achieved, but ambiguity and computational complexity increase

Engineering Contradiction:
Improvespatial state determination accuracyVSAvoidcomputational complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The body is segmented into multiple rigid body segments (e.g., head, torso, limbs) with defined joint connections. This segmentation allows the complex flexible body problem to be broken down into simpler rigid body pose estimations for each segment, reducing overall computational complexity while maintaining measurement precision through the structured relationship between segments.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

A body model with predefined joint constraints and segment relationships serves as an intermediary between the raw image data and the final pose estimation. This model mediates the reconstruction process by providing anatomical priors and constraints that reduce ambiguity in 3D coordinate reconstruction without requiring exhaustive computational search of all possible configurations.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Measurement precision

If complex models are used to reduce ambiguity in 3D reconstruction, then measurement precision improves, but device complexity increases

Engineering Contradiction:
Improvepose estimation accuracyVSAvoidmodel complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

Different levels of model complexity are applied to different body segments based on their specific requirements. Critical joints and segments that contribute most to pose accuracy receive more detailed modeling, while less critical areas use simplified representations. This local differentiation improves overall measurement precision without uniformly increasing model complexity across the entire body.

Inventive Principle:
Principle #3Local quality

Solution Approach 2:

The body model parameters (such as segment lengths, joint ranges of motion, and anatomical landmarks) are adapted and optimized for specific object classes (e.g., different human body types). This parameter customization improves measurement precision for specific applications without requiring fundamentally different complex models, thereby controlling overall model complexity while enhancing accuracy for targeted use cases.

Inventive Principle:
Principle #35Parameter changes

3Measurement precision

If methods are specialized for specific object classes, then measurement precision improves, but adaptability decreases

Engineering Contradiction:
Improvepose estimation accuracyVSAvoidobject class flexibility
Core Design Contradiction:
Measurement precisionVSAdaptability or versatility

Solution Approach 1:

The body model framework is designed with universal joint definitions and segment structures that can accommodate multiple object classes (e.g., different human body types, animals, or humanoid figures). By using a standardized hierarchical representation with configurable parameters, the system achieves high adaptability across different object classes while maintaining measurement precision through class-specific parameter optimization rather than requiring fundamentally different models for each class.

Inventive Principle:
Principle #6Universality (Multi-functionality)

Data Source

PatentEP3724852B1Method for the contactless determination of the degrees of freedom of a body
Publication Date: 2022.08.17 INFERICS GMBH
  • EP3724852B1 patent drawingFigure 1
  • EP3724852B1 patent drawingFigure 2~3

AI summary

The present invention relates to a method for the contactless determination of the internal and external degrees of freedom of a body in three-dimensional space. The internal degrees of freedom of the body are characterized by the positions of characteristic object points relative to one another or to a prominent object point. Alternatively, the internal degrees of freedom can be described by specifying the parameter values of a transformation which a body describes as an manifestation of a versatile object model. The values for the transformation parameters are estimated with the aid of 3D coordinates of characteristic objects of the body determined according to the invention, with the creation of a radiometric image and a depth image. The external degrees of freedom of the body describe the location and orientation thereof as a totality and are defined by the three-dimensional coordinates of a point of origin assigned to the body, such as the centre of gravity, and by the three-dimensional rotation of a coordinate system attached thereto in relation to an external coordinate system. The point of origin and the coordinate system attached thereto form the object coordinate system, the translation and rotation of which in relation to an external coordinate system represent the external degrees of freedom of the object.