Fiduciary Marker Tracking via Edge Detection and Geometric Modeling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Augmented reality systems face challenges in reliably estimating the position and orientation of fiduciary markers, especially when markers become distorted or too small due to camera angle or distance, leading to recognition failures during page turning in books and other objects.
Innovation Solution
The use of fiduciary markers with spatial redundancy and additional non-alphanumeric patterns on book pages, combined with an entertainment device's video camera and processing system, allows for detection and estimation of marker orientation and scale, even when markers are partially obscured, and employs alternative techniques to track page turning by hypothesizing edge positions and scoring criteria for accurate image augmentation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If fiduciary markers are used for object positioning in augmented reality, then the system can detect features and generate graphical overlays, but the markers become distorted or too small when viewed from certain angles or distances, leading to recognition failures
Solution Approach 1:
The system divides the tracking task into multiple components by using both fiduciary markers and non-fiduciary features (such as book spine edges, page edges, and corner detections). This segmentation allows the system to switch between different feature types depending on which are reliably detectable in the current view, maintaining tracking reliability even when markers become distorted or too small.
Solution Approach 2:
The system introduces an intermediary geometric model (book structure model with spine, pages, and corners) that mediates between the visual input and the tracking output. This model allows the system to infer marker position and orientation even when markers are not directly visible or are distorted, by using detectable features like book edges and corners as intermediaries to calculate the expected marker transformation.
2Adaptability or versatility
If the video camera captures the scene from various angles and distances, then the system can accommodate different viewing conditions, but the fiduciary marker quality degrades due to distortion and size reduction
Solution Approach 1:
The system dynamically adapts its feature detection strategy based on current viewing conditions. It continuously evaluates the quality and detectability of different feature types (fiduciary markers, non-fiduciary markers, book edges, corners) and switches between them in real-time. This dynamic adaptation allows the system to maintain reliable tracking across various angles and distances by always using the most suitable feature type for the current view.
Data Source
Figure 1~2A
Figure 2B
Figure 3
AI summary
An entertainment device comprises an input operable to receive a captured image from a video camera, a marker detector operable to detect a fiduciary marker within the captured image, and operable to estimate a distance and angle of the fiduciary marker, and a failure boundary calculation processor operable to calculate at least one of an additional distance and an additional angle from the currently estimated distance and angle of the fiduciary marker at which recognition of the fiduciary marker is assumed to fail.