Marker-Based Posture Reconstruction for Distal Joint Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for reconstructing body posture in virtual spaces face challenges with low recognition rates for distal body parts, often requiring motion sensors or depth sensors, which increase errors and complexity.
Innovation Solution
An electronic device identifies posture information by using markers attached to body parts, obtaining joint angles and positions through video analysis without separate motion or depth sensors, utilizing neural networks to connect joint information for accurate posture reconstruction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If motion sensors or depth sensors are used to accurately identify distal body parts, then measurement precision is improved, but device complexity increases
Solution Approach 1:
The patent introduces markers as intermediary objects attached to body parts, which serve as mediators between the camera and the body parts. These markers are easily detectable by the camera and provide clear reference points for identifying distal body parts like wrists and ankles, eliminating the need for complex motion sensors or depth sensors while maintaining high identification accuracy
Solution Approach 2:
The patent replaces the mechanical sensor-based detection system with an optical detection system using markers and standard cameras. This substitution eliminates complex mechanical sensors and depth sensors, reducing device complexity while maintaining measurement precision through visual marker detection and neural network processing
2Measurement precision
If motion sensors or depth sensors are used to accurately identify distal body parts, then measurement precision is improved, but error increases
Solution Approach 1:
The markers serve as reliable intermediary objects that provide clear, unambiguous detection targets. Their distinct visual characteristics ensure consistent and accurate detection by the camera, reducing errors compared to sensor-based methods that may suffer from calibration issues, noise, and environmental interference
Solution Approach 2:
The patent creates a virtual copy of the body posture by detecting marker positions and reconstructing the body model in virtual space. This copying approach allows for accurate representation of distal body parts without the errors inherent in direct sensor measurement, as the virtual model can be precisely calculated from multiple marker positions
Data Source
AI summary
An electronic device according to an embodiment includes a memory for storing instructions, and at least one processor, wherein when the instructions are executed, the at least one processor is configured to identify a video capturing a body and one or more markers attached to the body, obtain a first information including an angle and a position of at least one first joint among joints being included in the body, based on the one or more markers in the video, obtain a second information including a position of at least one second joint among the joints being included in the body, based on at least part of the video in which the body is captured among the body and the one or more markers in the video, and obtain a third information indicating a posture of the body based on the interconnection of the at least one first joint and the at least one second joint in the video, based on the first information and the second information.