AR Point-of-Interest Tracking Through Dual-Pipeline Feature Mapping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video object tracking techniques are limited by the use of a small number of trackable points, which compromises accuracy and quality in tracking objects, particularly facial features, and require significant computational resources and time.
Innovation Solution
A method and system that augment points of interest in augmented-reality videos by extracting feature points, determining a coordinate system, and using reference points to calculate and recalibrate location parameters, allowing for accurate tracking with minimal computational resources.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the number of trackable points is increased to improve tracking accuracy of object boundaries, then measurement precision is improved, but computational processing resources and time are significantly consumed
Solution Approach 1:
The patent segments the tracking process into two distinct pipelines: a first pipeline that processes video frames in real-time using a limited set of trackable points, and a second pipeline that performs computationally intensive processing to determine precise location parameters of points of interest. This segmentation allows the system to maintain real-time processing capability while achieving high tracking accuracy through the coordinated work of both pipelines.
Solution Approach 2:
The patent introduces an intermediary mechanism where the first pipeline extracts feature points and transfers them to the second pipeline, which then calculates precise location parameters and transfers them back to the first pipeline for augmentation. This intermediary exchange of data between pipelines enables the system to achieve high measurement precision without requiring the first pipeline to perform computationally expensive calculations directly.
2Measurement precision
If the number of trackable points is increased to accurately track outlines and boundaries of facial features, then measurement precision is improved, but device complexity increases
Solution Approach 1:
The patent divides the complex tracking system into two specialized pipelines with distinct functions. The first pipeline handles real-time video processing and feature extraction using a simple set of trackable points, while the second pipeline handles the complex calculations for determining precise location parameters. This segmentation reduces the complexity burden on any single component while achieving high overall precision.
Solution Approach 2:
The patent uses an intermediary data exchange mechanism between the two pipelines. The first pipeline transfers feature point data to the second pipeline, which then returns calculated location parameters. This intermediary approach allows the system to achieve accurate boundary tracking without requiring either pipeline to be overly complex, as each pipeline focuses on its specific task.
3Measurement precision
If location parameters are recalculated for every frame and time interval to maintain tracking accuracy, then measurement precision is improved, but loss of time increases
Solution Approach 1:
The patent performs preliminary action by having the second pipeline pre-calculate precise location parameters based on feature points extracted by the first pipeline. These pre-calculated parameters are then transferred back to the first pipeline for augmentation, eliminating the need for the first pipeline to perform time-consuming recalculations for every frame while maintaining high tracking accuracy.
Data Source
Figure 1~2B
Figure 2C~3
Figure 4
AI summary
A method for augmenting point of interest (202) in AR video. The method includes: extracting first frame (300) of AR video depicting object being tracked (200) including point of interest (202), by first pipeline; locating, in first frame (300) set of feature points forming boundary (206) of object; transferring frame (300) with determined set of feature points to second pipeline; and by second pipeline determining coordinate system (210) of set of feature points; calculating first location parameters of feature points; selecting first and second reference points (216) from feature points, wherein first location parameters of point of interest (202) are defined by first location parameters of first and second reference points (216); transmitting first location parameters of point of interest (202) from second pipeline to first pipeline; and augmenting in first pipeline AR video with first location parameters of point of interest (202) received from second pipeline.