3D Image Depth Adjustment via Region Correspondence Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing 3D image processing technologies lack user-friendly options for adjusting the depth display of objects within a 3D viewable synthetic image, requiring manual adjustments or fixed configurations, which limits user flexibility in focusing specific objects.
Innovation Solution
An image processing apparatus and method that allows users to select a region of interest, detect corresponding regions in other images with minimal parallax, and automatically adjust image positions to achieve optimal focus, using techniques like region assignment, correlation computation, and image recognition to form vivid 3D images with desired depth settings.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If the configuration in the depth direction is uniquely fixated according to well-known image processing technology, then the processing is simple and automatic, but the user has no margin to make any selection and cannot adjust which object is in focus
Solution Approach 1:
The system performs automatic region correspondence detection and parallax adjustment based on user's simple region assignment input. The computer automatically detects corresponding regions in other feed images, calculates parallax values, and adjusts image positions without requiring user expertise in depth configuration, making the system self-sufficient while providing user control.
Solution Approach 2:
The system pre-calculates and stores parallax information for multiple region combinations before final image synthesis. By preparing region correspondence data and parallax values in advance based on the user's assigned region, the system enables rapid focus adjustment without real-time complex computations during the actual depth modification process.
2Measurement precision
If manual operation is required to judge whether the relative position of the image is finely adjusted, then the focus accuracy can be improved, but the operation becomes complex and time-consuming
Solution Approach 1:
The system automatically calculates parallax values between the user-assigned region and corresponding regions, uses this feedback to determine optimal image position adjustments, and synthesizes the final image with precise focus control. This automated feedback loop eliminates the need for manual trial-and-error adjustment while maintaining high precision.
Solution Approach 2:
The patent replaces manual visual judgment and physical image positioning with automated computer-based parallax calculation and digital image processing. The system uses algorithmic parallax measurement and automated image registration to achieve precise focus adjustment without requiring manual intervention.
3Manufacturing precision
If the user has to carry out manual operation to judge fine adjustment of image relative position, then focus accuracy improves, but ease of operation deteriorates
Solution Approach 1:
The system automatically performs region correspondence detection, parallax calculation, and image position adjustment based on the user's simple region assignment. The computer serves itself by autonomously completing the complex alignment tasks without requiring user expertise or manual fine-tuning operations.
Solution Approach 2:
Manual visual inspection and physical image positioning are replaced with automated computer vision-based region detection and digital image registration algorithms. The system uses automated parallax measurement and computational methods to achieve precise alignment without manual intervention.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Enables users to easily adjust the depth display of images, ensuring objects of interest are in focus while maintaining vividness, without manual fine-tuning, by automatically processing the selected regions and their corresponding overlaps, enhancing user control over the 3D image appearance.
Implementation Method 1
a 3-D viewable synthetic image via a lenticular lens
Data Source
AI summary
To provide a solution by which adjustment of the depth display of the image can be easily carried out by the user at will in a technique for forming a 3-D image from plural images, from the plural feed images, one feed image is extracted as the reference feed image, with an object recognition process being carried out to extract the object region having the prescribed characteristic features. The reference feed image IL is displayed on the display unit 108 together with the markers MK indicating the object regions, and the user selects one object region. A region that is similar in image content with the selected region is detected from each other feed image, with the images being shifted so that the regions overlap each other.