3D Imaging Depth Resolution Control for Cardboard Effect Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional three-dimensional image capturing techniques face challenges in achieving compatibility between reducing calculation cost and improving stereoscopic effect, often resulting in a 'cardboard effect' due to static depth resolution and increased calculation costs.
Innovation Solution
A three-dimensional image capturing apparatus and method that dynamically adjusts depth values near designated objects, increasing depth resolution without increasing the overall number of depth values, thereby enhancing the stereoscopic effect while controlling calculation costs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the number of predetermined depth values is increased to improve depth resolution and three-dimensional appearance, then the stereoscopic effect is improved, but the calculation cost and memory requirements increase significantly
Solution Approach 1:
The patent applies local quality by setting different depth resolutions for different regions of the image. Specifically, depth values are densely set in regions containing objects of interest while using fewer depth values in background regions. This allows the system to improve depth resolution where it matters most (near objects) while avoiding the calculation cost increase that would result from uniformly high depth resolution across the entire image.
Solution Approach 2:
The patent implements dynamic depth value selection based on object detection results. The system first detects objects in the image, then dynamically adjusts the number and distribution of depth values according to the detected objects' positions and importance. This dynamic approach allows the system to optimize calculation resources by concentrating depth resolution on relevant regions rather than using a static, uniformly high depth resolution throughout.
2Measurement precision
If the number of predetermined depth values is increased to reduce the cardboard effect and improve three-dimensional appearance, then the visual quality is improved, but the memory required to hold calculation results increases
Solution Approach 1:
The patent reduces memory requirements by applying local quality to the depth value distribution. Instead of maintaining high depth resolution across the entire image which would require large memory, the system concentrates depth values in specific regions containing objects. This local concentration approach achieves improved three-dimensional appearance for important objects while significantly reducing the total number of depth values that must be stored in memory.
3Productivity
If static depth values are used across the entire image, then the calculation process is simple and fast, but the three-dimensional appearance is poor and cardboard effect occurs
Solution Approach 1:
The patent transitions from static to dynamic depth value selection. The system dynamically determines the number and distribution of depth values based on object detection results, allowing depth resolution to adapt to the content of each image region. This dynamic approach maintains processing efficiency by avoiding uniformly high depth resolution while still achieving good three-dimensional appearance through concentrated depth values in object regions.
Solution Approach 2:
The patent applies segmentation by dividing the image into different regions based on object detection results. Different depth resolution strategies are applied to different segments: high depth resolution is used in regions containing detected objects while lower depth resolution is used in background regions. This segmentation approach allows the system to optimize the balance between processing speed and three-dimensional appearance by treating different image regions differently.
Data Source
AI summary
A three-dimensional image capturing apparatus (100) generates depth information to be used for generating a three-dimensional image from an input image, and includes: a capturing unit (130) obtaining the input image in capturing; an object designating unit (110) designating an object in the input image; a resolution setting unit (120) setting depth values, each representing a different depth position, so that in a direction parallel to a depth direction of the input image, depth resolution near the object is higher than depth resolution positioned apart from the object, the object being designated by the object designating unit (110); and a depth map generating unit (140) generating two-dimensional depth information corresponding to the input image by determining, for each of regions in the input image, a depth value, from among the depth values set by the resolution setting unit (120), indicating a depth position corresponding to one of the regions.