Plenoptic Depth Estimation Using Pixel Homogeneity Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for estimating depth from plenoptic data are time-consuming, inaccurate, especially in non-textured areas, and suffer from low accuracy around edges and foreground fattening effects.
Innovation Solution
A method that estimates the inherent shift of in-focus pixels in focused plenoptic data by determining pixel homogeneity and excluding pixels with disparities equal to the inherent shift or belonging to homogeneous areas, using a combination of microlens grid distance, aperture, and focal length calculations, to improve disparity estimation accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If block-matching method is used for disparity estimation, then depth estimation can be performed, but accuracy is low around edges and in non-textured areas
Solution Approach 1:
The patent extracts and identifies pixels with homogeneous intensities across different views using a homogeneity metric. These homogeneous pixels are then excluded from disparity estimation, removing the source of error that causes edge accuracy degradation and foreground fattening effects.
Solution Approach 2:
The patent applies different processing treatments to different types of pixels based on their local properties. Textured pixels undergo standard disparity estimation while homogeneous pixels are excluded, creating a quality-aware processing approach that improves overall accuracy.
2Measurement precision
If structure tensors and eigenvalue decompositions are calculated for disparity estimation, then disparity can be estimated, but computational cost is high
Solution Approach 1:
The patent extracts homogeneous pixels using a simple intensity comparison metric and excludes them from further processing. This removes the need to perform computationally expensive structure tensor and eigenvalue decomposition calculations on pixels that would not contribute meaningful disparity information.
Solution Approach 2:
Instead of performing full disparity estimation on all pixels using computationally intensive methods, the patent applies a simplified homogeneity check to all pixels and only performs detailed disparity estimation on non-homogeneous pixels, reducing overall computational load while maintaining accuracy.
3Quantity of substance
If all pixels are used for disparity estimation, then more data is available for depth calculation, but accuracy is degraded by homogeneous areas
Solution Approach 1:
The patent identifies and extracts pixels with homogeneous intensities across different views using a homogeneity metric. These pixels are then removed from the set of pixels used for disparity estimation, preventing them from degrading the overall accuracy while retaining all useful textured pixels.
Solution Approach 2:
The patent evaluates each pixel's local intensity homogeneity and applies different processing strategies accordingly. Non-homogeneous pixels are used for disparity estimation while homogeneous pixels are excluded, creating a quality-based selection approach that improves precision.
Data Source
AI summary
Method and apparatus for estimating the depth of focused plenoptic data are suggested. The method includes: estimating the inherent shift of in-focus pixels of the focused plenoptic data; calculating a level of homogeneity of the pixels of the focused plenoptic data; determining the pixels of the focused plenoptic data which either have disparities equal to the inherent shift or belong to homogeneous areas, as a function of the level of homogeneity of the pixels of the focused plenoptic data; and estimating the depth of the focused plenoptic data by a disparity estimation without considering the determined pixels. According to the disclosure, the pixels of the focused plenoptic data which either have a disparity equal to the inherent shift or belong to a homogeneous area will not be considered for the estimation of the depth, which can reduce computational costs and at the same time increase accuracy of estimations for in-focus parts of the scene.