Rigged Smooth-Surface Model Tracker for Articulated Objects
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Articulated objects, such as the human body or a robot, are challenging to track accurately and efficiently from captured sensor data like video images and depth camera images due to the complexity of calculating global position and orientation, as well as joint positions.
Innovation Solution
A tracker that uses a processor to access a rigged smooth-surface model, computing pose parameters by optimizing the model fit to sensor data, with variables representing correspondences between the data and model included in the optimization process, enabling unified correspondence estimation and model fitting.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional tracking methods are used for articulated objects, then the tracking can be performed with existing algorithms, but the tracking accuracy and speed are insufficient
Solution Approach 1:
The patent transforms the tracking problem by changing the parameter representation from discrete joint angles to continuous pose parameters (position and orientation) of a rigged smooth-surface model. This parameter transformation enables the use of optimization algorithms that simultaneously estimate correspondences and pose parameters, achieving both high accuracy and speed by unifying the optimization process.
2Measurement precision
If complex optimization processes are used to improve tracking accuracy, then measurement precision improves, but computational time increases
Solution Approach 1:
The patent merges the correspondence estimation and pose parameter optimization into a single unified optimization process. By combining these two previously separate steps into one joint optimization, the system achieves accurate pose parameter estimation without the need for multiple sequential processing stages, thereby reducing overall computational time while maintaining high precision.
3Measurement precision
If separate processes are used for correspondence estimation and model fitting, then each process can be optimized independently, but the overall tracking accuracy suffers
Solution Approach 1:
The patent combines correspondence estimation and model fitting into a single unified optimization framework. This merging allows the system to achieve higher tracking accuracy by considering both correspondence and pose parameters simultaneously, while the modular structure of the unified optimizer keeps the implementation complexity manageable.
Data Source
AI summary
A tracker is described which comprises an input configured to receive captured sensor data depicting an object. The tracker has a processor configured to access a rigged, smooth-surface model of the object and to compute values of pose parameters of the model by calculating an optimization to fit the model to data related to the captured sensor data. Variables representing correspondences between the data and the model are included in the optimization jointly with the pose parameters.