Ground Truth Dataset Generation for Realistic VR Motion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current virtual reality (VR) systems generate synthetic and artificially smooth trajectories, resulting in unrealistic user movements, such as virtual cameras moving like aircraft rather than humans walking or running, due to the limitation of using tailored datasets for specific headsets and environments with limited motion primitives.
Innovation Solution
A method involving a wearable device with cameras and inertial measurement units to record real motion data, which is used to generate a continuous time trajectory function. This function is then applied to simulate sensors in a virtual environment, creating a ground truth dataset that replicates realistic user movements, allowing for the generation of multiple datasets for testing various VR headset configurations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If synthetic and artificially smooth trajectories are used in VR systems, then the system operation is simplified and device complexity is reduced, but the realism of user movements deteriorates
Solution Approach 1:
The patent copies real human motion data captured from wearable devices and applies it to virtual camera trajectories. Instead of generating synthetic smooth trajectories, the system records actual human movements including natural variations and applies them to VR experiences, preserving realism while simplifying the trajectory generation process.
Solution Approach 2:
The patent performs preliminary recording of real motion data from wearable devices before VR experiences. The captured motion data is stored and reused across multiple VR sessions and configurations, eliminating the need to generate synthetic trajectories each time and maintaining consistent realistic movement patterns.
2Measurement precision
If tailored datasets are used for specific headsets and environments, then the accuracy for specific applications is improved, but the adaptability to different configurations deteriorates
Solution Approach 1:
The patent creates motion datasets that are universally applicable across different VR headset configurations and environments. By capturing generic human motion patterns rather than device-specific trajectories, the same dataset can be applied to multiple headset types, resolutions, and display configurations, eliminating the need for separate tailored datasets.
Solution Approach 2:
The system copies real human motion data that inherently captures universal movement patterns applicable across different devices. These copied motion patterns serve as transferable assets that maintain accuracy across various VR configurations without requiring re-generation for each specific headset.
3Device complexity
If limited motion primitives are used, then the complexity of motion generation is reduced, but the realism and variety of user movements deteriorates
Solution Approach 1:
The patent copies complete, complex real human motion sequences directly from wearable device recordings rather than constructing movements from limited primitives. This approach captures the full complexity and variety of natural human movement including subtle variations, accelerations, and decelerations that would be difficult to generate from simple motion primitives.
Solution Approach 2:
The system uses dynamic, continuously recorded motion data from wearable devices rather than static motion primitives. The captured data includes real-time variations in speed, direction, and movement patterns that naturally provide variety and realism without requiring complex generation algorithms.
Data Source
AI summary
Systems and methods of generating ground truth datasets for producing virtual reality (VR) experiences, for testing simulated sensor configurations, and for training machine-learning algorithms. In one example, a recording device with one or more cameras and one or more inertial measurement units captures images and motion data along a real path through a physical environment. A SLAM application uses the captured data to calculate the trajectory of the recording device. A polynomial interpolation module uses Chebyshev polynomials to generate a continuous time trajectory (CTT) function. The method includes identifying a virtual environment and assembling a simulated sensor configuration, such as a VR headset. Using the CTT function, the method includes generating a ground truth output dataset that represents the simulated sensor configuration in motion along a virtual path through the virtual environment. The virtual path is closely correlated with the motion along the real path as captured by the recording device. Accordingly, the output dataset produces a realistic and life-like VR experience. In addition, the methods described can be used to generate multiple output datasets, at various sample rates, which are useful for training the machine-learning algorithms which are part of many VR systems.