3D Shape Reconstruction Using Surface Data to Refine Visual Hulls
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for generating three-dimensional shape data, such as the visual hull method, often result in inaccuracies when dealing with objects that have curved or concave surfaces, leading to errors in the approximation of these shapes.
Innovation Solution
A method involving multiple cameras capturing images from different viewpoints to generate approximate shape data, followed by correcting the distance image using surface three-dimensional information and setting threshold values to refine the shape data, ensuring accuracy even for complex shapes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If the visual hull method is used to generate three-dimensional shape data, then the processing speed and stability are improved, but the measurement precision deteriorates due to errors in approximating curved and concave surfaces
Solution Approach 1:
The patent segments the three-dimensional shape restoration process into two distinct stages: first generating an approximate shape model using the visual hull method, then refining this model by restoring accurate shapes of concave portions using silhouette images and local shape functions. This segmentation allows each stage to optimize for its specific purpose while achieving overall high accuracy.
Solution Approach 2:
The patent performs preliminary action by first generating an approximate shape model that serves as a foundation for subsequent refinement. This preliminary model provides initial three-dimensional information that guides the later restoration process, making the overall system more efficient and accurate.
2Device complexity
If the visual hull method approximates surfaces as flat surfaces, then the device complexity is reduced, but the manufacturing precision deteriorates due to inability to represent curved and concave surfaces accurately
Solution Approach 1:
The patent segments the shape representation into two parts: flat surface approximation for overall structure (visual hull) and curved/concave surface restoration for detailed accuracy. This segmentation allows the system to maintain low complexity for the base model while achieving high precision through targeted refinement of specific surface regions.
Solution Approach 2:
The patent applies local quality by restoring accurate local shapes only in regions where concave portions exist, rather than applying complex modeling techniques globally. This is achieved by using silhouette images and local shape functions specifically for concave region restoration, maintaining simplicity elsewhere.
3Measurement precision
If local shape information is restored based on approximate shape models, then the measurement precision is improved, but the reliability deteriorates when the difference between approximate and actual local shapes cannot be compensated
Solution Approach 1:
The patent implements feedback by using silhouette images as verification against the restored three-dimensional shape. The silhouette information provides continuous feedback to identify regions where restoration is insufficient, allowing the system to iteratively improve accuracy and maintain reliability by compensating for errors in distance information.
4Measurement precision
If multiple procedures are used to restore accurate three-dimensional shape, then the measurement precision is improved, but the productivity decreases due to increased processing steps
Solution Approach 1:
The patent segments the processing into distinct functional stages (visual hull generation, silhouette extraction, local shape restoration) that can be executed efficiently in sequence. Each segment is optimized for its specific task, preventing redundant computations and maintaining productivity despite the multi-step process.
Solution Approach 2:
The patent performs preliminary action by generating the approximate shape model first, which serves as a foundation that reduces the complexity of subsequent restoration steps. This preliminary structure allows later procedures to focus only on refining specific regions rather than processing the entire shape from scratch.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
An object is to obtain three-dimensional shape data with high accuracy from three-dimensional shape data representing an approximate shape of an object. Three-dimensional shape data of an object captured in a plurality of captured images whose viewpoints are different is obtained (302). Further, surface three-dimensional information on the object is derived (303) based on the plurality of captured images. Then, the derived surface three-dimensional information is selected (304) based on a distance from the shape surface of the object represented by the three-dimensional shape data.