Super-Resolution Mask Expansion for 3D Shape Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing technique for generating three-dimensional shape data using the visual hull method suffers from reduced accuracy due to low-resolution mask images, which can omit parts of the object, leading to incomplete three-dimensional shape data.
Innovation Solution
An image processing system that identifies low-resolution areas in captured images and performs expansion processing on foreground masks within these areas to enhance the resolution, ensuring more accurate three-dimensional shape data generation using the visual hull method.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If mask images with low resolution are used as extracted foregrounds, then the processing speed is improved, but the accuracy of three-dimensional shape data is reduced
Solution Approach 1:
The patent applies local quality by performing super-resolution processing selectively only on mask images with low resolution rather than all mask images. The processing unit identifies which mask images require enhancement based on their resolution characteristics, applying computational resources only where needed to maintain accuracy while preserving processing efficiency for already-high-resolution images.
Solution Approach 2:
The patent changes the resolution parameter of mask images through super-resolution processing. By transforming low-resolution mask images into high-resolution versions, the system modifies the key parameter (resolution) to improve the accuracy of three-dimensional shape data generation without fundamentally changing the visual hull method workflow.
2Device complexity
If mask images with low resolution are used as extracted foregrounds, then the device complexity is reduced, but the accuracy of three-dimensional shape data is reduced
Solution Approach 1:
The system implements local quality by applying super-resolution processing only to specific mask images that have low resolution, rather than uniformly processing all images or upgrading the entire system hardware. This targeted approach maintains simplicity in the overall device architecture while locally enhancing where needed.
Solution Approach 2:
The patent replaces a mechanical/hardware solution (upgrading imaging devices to capture higher resolution images) with a computational solution (super-resolution processing). This substitution maintains the existing device configuration while achieving improved accuracy through image processing algorithms that reconstruct high-resolution details from low-resolution inputs.
3Measurement precision
If super-resolution processing is performed on mask images, then the accuracy of three-dimensional shape data is improved, but the processing time is increased
Solution Approach 1:
The patent minimizes time loss by applying super-resolution processing only to mask images that have low resolution and would benefit from enhancement. Mask images that are already high-resolution skip this processing step, thereby reducing the total processing time while still improving accuracy where it matters most.
Solution Approach 2:
The system performs partial action by selectively applying super-resolution processing only to the extent necessary - specifically to low-resolution mask images that would result in inaccurate three-dimensional shape data. This avoids the excessive processing time that would result from applying the same processing to all mask images regardless of their initial quality.
Data Source
AI summary
The image processing apparatus obtains an image representing an object, performs specific processing for an object included in an area whose object extraction accuracy is lower than a threshold value in the captured image, and generates three-dimensional shape data of the object based on the image representing the object for which the specific processing has been performed.