Predictive Model Updating for 3D Gesture Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional motion capture systems are expensive, cumbersome, and often operate in real-time due to the complexity of data analysis, limiting their deployment and use.
Innovation Solution
The technology simplifies the updating of a predictive model by clustering observed points in a 3D sensory space, refining positions of model segments, and matching clusters to improve predictive information for real-time control and communication.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional motion capture systems use markers or sensors worn by subjects and strategic placement of numerous cameras, then measurement precision can be improved, but device complexity and cost increase significantly
Solution Approach 1:
The patent extracts and removes the need for physical markers and sensors from the motion capture system. Instead of requiring subjects to wear markers or sensors, the system uses computer vision algorithms to detect and track natural body landmarks (joints, limbs) directly from video footage, eliminating the complexity of marker attachment and sensor equipment while maintaining measurement capability
Solution Approach 2:
The patent creates a virtual 3D model (skeletal model) that copies and represents the subject's body structure. This digital twin allows the system to track motion by matching observed video data against the predefined skeletal model, enabling accurate motion capture without physical markers by reconstructing the subject's pose from 2D video images into 3D space
2Measurement precision
If conventional motion capture systems place numerous cameras strategically to capture subject movements, then measurement precision improves, but ease of operation deteriorates due to cumbersome setup
Solution Approach 1:
The patent makes the system universal by enabling it to work with standard video cameras and video input devices that are already ubiquitous. The same system can track multiple subjects simultaneously and adapt to different environments without requiring specialized camera placements, making it as easy to operate as any video input device while maintaining tracking accuracy
Solution Approach 2:
The system performs automatic calibration and subject identification without requiring manual setup. The computer vision algorithms automatically detect subjects, identify body landmarks, and establish tracking parameters from the video feed itself, eliminating the need for operators to perform complex camera calibration procedures or manually configure the system for each subject
3Measurement precision
If conventional motion capture systems analyze and correlate large volumes of data from multiple cameras, then measurement precision improves, but productivity decreases due to real-time processing requirements
Solution Approach 1:
The patent segments the motion capture problem into independent tracking of individual body landmarks rather than analyzing all camera data simultaneously. The system identifies and tracks specific anatomical points (joints, limbs) as separate entities, allowing parallel processing of each landmark's trajectory independently, which reduces computational complexity and enables real-time performance
Solution Approach 2:
The system performs preliminary action by pre-defining the skeletal model with known anatomical relationships and constraints before tracking begins. This predefined structure allows the algorithms to make intelligent predictions about landmark positions and reduce the search space during real-time processing, significantly speeding up computation while maintaining accuracy
Data Source
AI summary
The technology disclosed relates to simplifying updating of a predictive model using clustering observed points. In particular, it relates to observing a set of points in 3D sensory space, determining surface normal directions from the points, clustering the points by their surface normal directions and adjacency, accessing a predictive model of a hand, refining positions of segments of the predictive model, matching the clusters of the points to the segments, and using the matched clusters to refine the positions of the matched segments. It also relates to distinguishing between alternative motions between two observed locations of a control object in a 3D sensory space by accessing first and second positions of a segment of a predictive model of a control object such that motion between the first position and the second position was at least partially occluded from observation in a 3D sensory space.