Intermediate View Image Generation Using Selective Disparity Calculation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for generating intermediate view images in stereoscopic images face challenges such as requiring uniform disparity information, which limits their use to simple images, or calculating disparity information for all pixels, which is time-consuming and prone to noise, leading to poor quality intermediate views.
Innovation Solution
An apparatus and method that detect feature points and corresponding points using filtering techniques like Canny and Sobel filtering, and then calculate disparity information selectively, focusing on important contours and removing abnormal corresponding points to generate high-quality intermediate view images.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If disparity information is calculated for all pixels of the image, then the intermediate view can be obtained for complex images, but the calculation time becomes very long and the quality depends on the accuracy of the disparity map
Solution Approach 1:
The patent divides the image into multiple regions and performs disparity calculation selectively on important regions rather than all pixels. This segmentation approach maintains the ability to handle complex images while significantly reducing the overall calculation time by focusing computational resources on critical areas.
Solution Approach 2:
The patent applies different processing quality levels to different regions of the image. Important regions receive full disparity calculation while less important regions use simplified methods or uniform disparity values, optimizing the balance between image quality and processing time.
2Reliability
If disparity information is calculated for all pixels, then complete coverage is achieved, but noise and incorrect disparity information are easily generated
Solution Approach 1:
The patent extracts and removes abnormal corresponding points and incorrect disparity information from the disparity map using filtering techniques. This extraction of harmful elements reduces noise and improves the reliability of the final intermediate view image.
Solution Approach 2:
The patent implements feedback mechanisms where the disparity map is continuously refined by detecting and removing abnormal points. The system uses the generated intermediate view to validate disparity information and iteratively improves accuracy by eliminating incorrect data.
3Productivity
If uniform disparity information is used for all regions, then the processing is simple and fast, but it can only be used for images without complicated changes
Solution Approach 1:
The patent applies uniform disparity information to some regions while performing detailed disparity calculation on other regions. This local differentiation allows the system to maintain fast processing for simple areas while accurately handling complex regions with significant depth variations.
Solution Approach 2:
The patent segments the image into regions requiring uniform disparity treatment and regions requiring detailed disparity calculation based on complexity metrics. This segmentation enables the system to optimize processing speed while maintaining adaptability to complex image structures.
Data Source
AI summary
Disclosed herein are an apparatus and a method for generating an intermediate view image of a stereoscopic image. The apparatus may include a feature point detector detecting contours in left and right view images and detecting feature points in the contours, a corresponding point detector detecting corresponding points corresponding to the feature points of the left and right view images, and a composer generating an intermediate view image based on disparity information between the feature points and the corresponding points.