3D Head Position Estimation Using Silhouette Contours
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing head tracking techniques are computationally expensive, require external markings or specialized equipment, and are vulnerable to changes in lighting and inaccuracies in 3D model formation, making them non-robust and unable to operate in real-time.
Innovation Solution
A method and system that uses calibrated cameras to capture images from different viewpoints, extracts silhouettes, and determines the 3D position of a head using elliptical models based on contours, allowing for real-time tracking without external markings or specialized gear, and is invariant to lighting changes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If facial features are extracted from images to track head position, then tracking reliability is improved, but computational cost increases significantly
Solution Approach 1:
The patent extracts only the essential silhouette contour of the head from the image, removing unnecessary internal facial features and details. This extraction of the boundary contour alone provides sufficient information for 3D position estimation while dramatically reducing computational complexity compared to full facial feature extraction.
Solution Approach 2:
Instead of extracting internal facial features as conventional methods do, the patent inverts the approach by using the external silhouette contour as the primary tracking feature. This inversion leads to lower computational cost while maintaining tracking reliability through the use of calibrated multi-view geometry.
2Measurement precision
If explicit 3D model of the user's head is generated to track position, then tracking accuracy is improved, but computational cost and time increase
Solution Approach 1:
The patent applies partial action by generating only the necessary 3D information from silhouette contours rather than creating a complete explicit 3D model of the head. The method computes 3D position estimates from the contours of multiple views using calibrated camera geometry, providing sufficient accuracy for tracking without the excessive computational burden of full 3D reconstruction.
3Speed
If optical flow or texture registration is computed to track head position, then motion tracking is achieved, but the system becomes vulnerable to lighting changes
Solution Approach 1:
The patent converts the potential harm of lighting changes affecting internal facial features into a benefit by relying exclusively on the silhouette contour. The contour is extracted from the boundary of the head in the image, which remains stable and unchanged under varying lighting conditions, thus eliminating the vulnerability to lighting changes while maintaining tracking speed.
4Reliability
If markings or LED lights are placed on the user's head for tracking, then tracking reliability is improved, but user comfort and ease of operation deteriorate
Solution Approach 1:
The patent applies self-service by using the user's own natural head silhouette contour as the tracking feature, without requiring any external markings, attachments, or specialized equipment. The system processes the user's appearance as it naturally occurs in the image, making the tracking process comfortable and convenient for the user while maintaining reliability through multi-view calibrated geometry.
Data Source
AI summary
A method and system for estimating the three dimensional position of an object in a three dimensional physical space. Specifically, the method discloses capturing a plurality of images of a human form within the three dimensional (3D) physical space. Each of the plurality of images is captured from a different viewpoint location of the human form. At least one image capturing device calibrated within the 3D physical space is used to capture the images. A plurality of silhouettes of the human form is extracted from the plurality of images. A plurality of contours of an object of the human form is obtained from the plurality of silhouettes. A location of the object within the 3D physical space is determined from an object model of the object based on the plurality of contours.