Body Tracking Estimation Using Server-Based Machine Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current body tracking techniques require users to wear trackers on the chest or waist, which is cumbersome, and struggle to accurately estimate the position of joints like the elbow without corresponding wrist orientation data, leading to inaccurate estimations during motions like hand waves.
Innovation Solution
An estimation apparatus and method that uses time-series data from trackers on the head, hands, and feet to estimate the orientation or angular velocity of the chest or waist, and the orientation of wrists, employing machine learning models to process data on orientations, angular velocities, and positions, allowing for accurate body tracking without the need for trackers on the chest, waist, or wrists.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If trackers are worn on the chest or waist to achieve accurate body tracking, then measurement precision is improved, but ease of operation deteriorates due to the troublesome wearing process
Solution Approach 1:
The patent extracts the estimation function from the tracker hardware itself and relocates it to a server-based machine learning system. Trackers only need to collect basic motion data from joints they are attached to, while the server performs complex inverse kinematics calculations to estimate positions of untracked body parts, eliminating the need for multiple trackers on the chest/waist
Solution Approach 2:
The patent introduces a server as an intermediary between the trackers and the body tracking system. This server receives data from a minimal set of trackers (ideally just one on the chest/waist) and uses machine learning models to infer and estimate the positions and orientations of multiple body parts, acting as a computational mediator that replaces the need for direct tracking of all body parts
2Measurement precision
If trackers are worn on the wrist to estimate elbow position, then measurement precision is improved, but ease of operation deteriorates due to the troublesome wearing process
Solution Approach 1:
The patent extracts the complex calculation function from the wrist tracker and relocates it to the server. The wrist tracker only needs to provide basic orientation data, while the server performs inverse kinematics calculations to estimate elbow position and other body part positions, eliminating the need for the user to wear multiple trackers
Solution Approach 2:
The patent creates a virtual copy of the body tracking system through machine learning models that replicate the functionality of multiple physical trackers. The server runs simulations and calculations that copy the behavior of what would occur if multiple trackers were worn, providing the same information with fewer physical devices
3Measurement precision
If multiple trackers are worn on different body parts to improve estimation accuracy, then measurement precision is improved, but device complexity increases
Solution Approach 1:
The patent makes the server system universal by implementing machine learning models that can estimate multiple different body part positions and orientations from a single set of input tracker data. The same server infrastructure handles estimation for chest, waist, elbows, wrists, and other body parts, making the system multi-functional rather than requiring separate tracking solutions for each body part
Solution Approach 2:
The server acts as an intermediary that consolidates data from minimal trackers and performs comprehensive body tracking calculations. Instead of having multiple trackers directly providing data for multiple body parts, the server mediates this relationship by receiving limited input data and generating comprehensive output estimates through machine learning models
Data Source
AI summary
Provided are a body part orientation estimation apparatus, a body part orientation estimation method, and a program that enable accurate body tracking without having the user wear many trackers. A time-series data input section (68) acquires a plurality of pieces of time-series data each representing positions, postures, or motions of a part of a body. The time-series data input section (68) inputs the plurality of pieces of time-series data into a conversion section (60). An output acquisition section (70) acquires a result of estimation of a position, a posture, or a motion of another part of the body that is closer to a center of the body than the part, the result of the estimation being an output obtained when the pieces of time-series data are input into the conversion section (60).