Multi-View Body-Part Tracking with Visibility-Weighted 3D Localization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing technologies face challenges in accurately identifying the locations of body parts in three-dimensional space from multiple images captured from different viewpoints, particularly when occlusions occur, leading to inaccuracies in tracking and representation.
Innovation Solution
An electronic device employs a neural network to process images from different viewpoints, adjusting weights based on visibility values and occlusions to track body part locations in a virtual three-dimensional space, using a combination of two-dimensional and three-dimensional representations to enhance accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If body part locations are identified from multiple images captured from different viewpoints, then the completeness of body part detection is improved, but the accuracy deteriorates when occlusions occur
Solution Approach 1:
The patent changes the parameter of weight assignment based on visibility values. Each body part location is assigned a weight according to its visibility in the image, allowing the system to adjust the importance of different detections dynamically. This resolves the contradiction by ensuring that visible body parts contribute more to the final result while reducing the impact of occluded parts.
Solution Approach 2:
The system uses visibility values as feedback to adjust the weights of body part locations. By continuously evaluating how much of each body part is visible and using this information to modulate detection confidence, the system can adapt to occlusion conditions and maintain accuracy while preserving detection completeness.
2Measurement precision
If visibility values are used to adjust location weights, then the accuracy of body part location tracking is improved, but the device complexity increases
Solution Approach 1:
The patent segments the processing into distinct modules: image input, body part detection, visibility value calculation, weight assignment, and location tracking. This segmentation allows each component to handle a specific task independently, reducing overall system complexity while maintaining high accuracy through coordinated operation of these modular components.
Data Source
AI summary
An electronic device includes memory storing instructions; and one or more processors, wherein the instructions, when executed by the one or more processors, cause the electronic device to input, into a neural network, at least one image, from among images that are obtained from different viewpoints, to obtain first information indicating locations of body parts of a subject included in the images, wherein the first information includes first location data indicating a first locations of portions of the body parts within a first image from among the images, and wherein the first locations are determined based on a first visibility values of the portions in the first image; obtain, based on the first information, second information indicating a second locations of the body parts at moments when the images were obtained; and track positions of the body parts in a virtual three-dimensional space based on the second information.