3D Photo Creation System Using Optical Flow Depth Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current 3D photo creation methods are inefficient, requiring extensive manual processing time and resulting in poor photo quality, especially when dealing with subjects in motion or multi-view angle image capture.
Innovation Solution
A 3D photo creation system and method that includes a stereo image input module, depth estimation module, multi-view angle image reconstructing module, and image spaced scanning module, which automatically estimates depth information, creates multi-view angle images, and adjusts them to form a mixed image, using pixel matching, optical flow calculation, and interpolation to enhance image quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual operation is used to convert 2D image into multi-view angle images, then depth information can be assigned based on operator judgment, but processing time increases to several hours to several days
Solution Approach 1:
The patent replaces manual mechanical operations with an automated computer-based system that uses optical flow algorithms and pixel matching to automatically extract depth information from stereo images, eliminating the need for manual mask creation and depth assignment while maintaining accuracy
Solution Approach 2:
The system enables self-service by automatically processing stereo images through programmed algorithms that perform pixel comparison, optical flow calculation, and depth map generation without requiring operator intervention, allowing the computer to independently complete the entire multi-view angle image creation process
2Adaptability or versatility
If multiple cameras are set up to photo-shoot subjects from multi-view angles, then multi-view angle images can be captured, but the method is not feasible when applying to subjects in motion and requires precise positioning
Solution Approach 1:
The patent creates virtual copies of view angles by using a single stereo image pair and generating multiple intermediate views through computational methods, eliminating the need for multiple physical cameras while achieving the same multi-view angle effect
Solution Approach 2:
The system transitions from spatial dimension (multiple cameras positioned at different locations) to computational dimension (algorithmic generation of multiple views from a single stereo pair), adding a new dimension of flexibility that allows motion capture without physical repositioning
3Productivity
If multi-view angle images are constructed directly from captured data, then the mixed image can be created, but the image quality cannot preserve the quality of the original image since it is a sub-sample
Solution Approach 1:
The patent replaces direct mechanical sampling with computational image processing techniques including optical flow-based pixel mapping and interpolation algorithms that reconstruct high-quality multi-view images by calculating pixel positions and intensities rather than simply subsampling captured data
Data Source
AI summary
The present application is directed to a 3D photo creation system and method, wherein the 3D photo creation system including: a stereo image input module configured to input a stereo image; wherein the stereo image comprises a left eye image and a right eye image; a depth estimation module configured to estimate a depth information of the stereo image and create a depthmap; a multi-view angle image reconstructing module configured to create a multi-view angle image according to the depthmap and the stereo image; and an image spaced scanning module configured to adjust the multi-view angle image and form a mixed image. The system and method outstandingly simplified the process of 3D photo creation and enhanced the quality of 3D photo. The system and method can be widely used in various theme parks, tourists attraction spots and photo galleries and bring about pleasure to more consumers with the 3D photos.