Auto-stereoscopic Interpolation Using Color Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for generating 3D images from 2D images are time-consuming and inefficient due to the complexity of accurately extracting objects and creating depth information, particularly in auto-stereoscopic imaging which requires precise selection of common points between images.
Innovation Solution
A computerized method and system that employs color segmentation to generate color segmented frames, calculates keys for each frame, and creates a depth map with three-dimensional information for each pixel, allowing for efficient and accurate 3D conversion by automatically defining object boundaries and adjusting Bezier points for improved detail.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If color segmentation is used to extract objects from 2D images, then object extraction accuracy is improved, but processing time increases
Solution Approach 1:
The patent applies color segmentation to divide the 2D image into multiple color-based regions or segments. This segmentation allows the system to identify and extract objects based on their color characteristics, improving object extraction accuracy while managing processing time through efficient segmentation algorithms.
Solution Approach 2:
The system dynamically adjusts segmentation parameters and depth map generation parameters based on image characteristics and processing requirements. By optimizing these parameters, the system achieves high extraction accuracy while controlling processing time through adaptive parameter adjustment.
2Measurement precision
If depth maps are generated using stereo pairs with common point correlation, then depth information accuracy is improved, but processing complexity and time increase
Solution Approach 1:
The patent extracts and utilizes only the essential common points between stereo images for depth map generation, rather than processing all image data. This extraction approach maintains depth information accuracy while significantly reducing processing complexity and time requirements.
Solution Approach 2:
The system performs partial depth map generation by focusing on critical regions and common points rather than complete image processing. This partial action approach achieves sufficient depth accuracy without the full processing complexity of traditional stereo pair methods.
3Productivity
If Bezier curves are used to define object outlines, then processing efficiency is improved, but outline accuracy decreases
Solution Approach 1:
The patent dynamically adjusts the number and positioning of Bezier curve points based on object complexity and required accuracy. For simple objects, fewer points are used for efficiency, while complex objects receive more points for accuracy, creating a dynamic balance between processing efficiency and outline accuracy.
Solution Approach 2:
The system applies different levels of Bezier curve detail to different regions of the image based on local object characteristics. High-detail curves are applied where needed for accuracy, while simplified curves are used where efficiency is prioritized, achieving local optimization of both productivity and precision.
Data Source
AI summary
Described are computer-based methods and apparatuses, including computer program products, for auto-stereoscopic interpolation. A first two dimensional image and a second two dimensional image are received. A reduced pixel image is generated for each of the first and second two dimensional images, wherein each reduced pixel image comprises a reduced pixel size that is less than the original pixel size. Boundary information is calculated for each of the first and second two dimensional images. A depth map is calculated for the first and second reduced pixel images, wherein the depth map comprises data indicative of three dimensional information for one or more objects in the first and second reduced pixel images. A depth map is calculated for the first and second two dimensional images based on the boundary information for each of the first and second two dimensional images and the depth map of the first and second reduced pixel images.