Markerless Motion Capture via Inverse Kinematics Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current motion capture technologies face challenges in achieving high accuracy for markerless motion measurement, particularly outdoors or in wide spaces, due to limitations in temporal and spatial resolution, and require extensive preparation and equipment, which is costly and labor-intensive.
Innovation Solution
A method for obtaining joint positions of a subject using a single input image or multiple images, involving spatial distribution likelihood calculation, optimization through inverse kinematics, and forward kinematics to estimate joint angles and positions, allowing for 3D reconstruction and smooth motion measurement without markers, regardless of indoor or outdoor environments.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If optical markers or sensors are attached to the subject body for motion capture, then measurement precision is improved, but device complexity and preparation time increase
Solution Approach 1:
The patent extracts the motion capture function from physical markers/sensors and implements it through software-based detection of natural body features in images. The system detects joints and body parts directly from image data without requiring additional physical attachments, thereby eliminating the complexity of marker attachment while maintaining measurement capability
Solution Approach 2:
The patent creates a virtual 3D model that copies the subject's body structure and motion characteristics. By constructing a virtual articulated model from 2D image observations and simulating its motion, the system achieves accurate motion capture without physical markers, replacing the need for complex physical sensing equipment with computational modeling
2Measurement precision
If multiple cameras are used for optical motion capture, then measurement precision is improved, but loss of time and loss of substance increase
Solution Approach 1:
The patent makes a single camera perform multiple functions: it captures images for both visual record and motion capture analysis. The same image used for viewing also serves as the basis for detecting joint positions and calculating 3D motion, eliminating the need for separate marker-based systems and reducing preparation time
Solution Approach 2:
The system creates a virtual 3D representation from 2D images, copying the essential motion information needed for analysis. This virtual model allows accurate motion reconstruction without requiring multiple physical cameras or extensive setup time
3Measurement precision
If optical markers are attached to the subject body, then measurement precision is improved, but ease of operation deteriorates due to constraint of movement
Solution Approach 1:
The patent removes the physical constraints of markers and sensors from the subject body. By detecting natural body features (joints, contours) directly from images, the system maintains measurement precision while allowing completely natural movement without any physical attachments that could restrict motion
4Device complexity
If deep learning techniques are used for markerless motion capture, then device complexity is reduced, but measurement precision may deteriorate
Solution Approach 1:
The patent segments the motion capture process into distinct computational stages: image acquisition, joint detection through deep learning, 3D coordinate calculation, and motion reconstruction. This segmentation allows each stage to be optimized independently, with deep learning handling detection and mathematical models handling precise 3D reconstruction, thereby maintaining accuracy while simplifying the overall system
Solution Approach 2:
The patent introduces a virtual articulated 3D model as an intermediary between 2D image observations and 3D motion reconstruction. This virtual model serves as a mediator that incorporates anatomical constraints and motion physics, ensuring measurement precision is maintained even though the physical system is simplified to a single camera
Data Source
AI summary
The present invention provides a motion capture with a high accuracy which can replace an optical motion capture technology, without attaching optical markers and sensors to a subject. A subject with an articulated structure has a plurality of feature points in the body of the subject including a plurality of joints wherein a distance between adjacent feature points is obtained as a constant. A spatial distribution of a likelihood of a position of each feature point is obtained based on a single input image or a plurality of input images taken at the same time. One or a plurality of position candidates corresponding to each feature point are obtained based on the spatial distribution of the likelihood of the position of each feature point. Each join angle is obtained by performing an optimization calculation based on inverse kinematics using the candidates and the articulated structure. Positions of the feature points including the joints are obtained by performing a forward kinematics calculation using the joint angles.