Disparity Map Correction via Local Neighborhood Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Disparity maps obtained from stereo matching algorithms or depth maps from hardware often contain errors that can hinder algorithm performance, with conventional methods like low-pass filtering or segmenting being inadequate for large structured mistakes and blurring edges.
Innovation Solution
A disparity map correction technique using statistical analysis on local neighborhoods to detect and correct outliers by segmenting images into regions of similar color, performing statistical analysis to identify and replace invalid disparities with values from neighboring pixels, and iteratively refining the results.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a low-pass filter is used to remove errors in disparity maps, then noise is reduced, but edges are blurred and high-frequency features are lost
Solution Approach 1:
The image is divided into multiple local neighborhoods or segments based on color homogeneity. This segmentation allows the algorithm to process different regions independently, preserving edges between segments while applying error correction within each segment. The segmentation prevents blurring across boundaries that would occur with global low-pass filtering.
Solution Approach 2:
The algorithm applies different processing strategies to different local regions. Within each segment, statistical analysis is performed to identify and correct errors based on local disparity distributions. This local approach maintains high-frequency features and edges while still removing noise, as each region is processed according to its own characteristics rather than applying a uniform blur across the entire image.
2Reliability
If the image is segmented into homogeneous regions and small regions are removed, then small homogeneous mistakes are corrected, but large structured mistakes and non-homogeneous regions are not addressed
Solution Approach 1:
The algorithm performs multiple iterations of disparity map correction with varying segmentations. In each iteration, the image is re-segmented and statistical analysis is performed with different parameters. This allows the system to adapt to different error types and scales, correcting both small homogeneous mistakes and larger structured errors by adjusting the segmentation and statistical thresholds across iterations.
Solution Approach 2:
The algorithm uses iterative refinement where the output of one iteration becomes the input for the next. Statistical analysis identifies outliers and corrects them, then the process repeats with updated disparity values. This feedback loop allows the system to progressively correct various types of errors, including large structured mistakes that may not be apparent in a single pass.
3Manufacturing precision
If planes are fitted to segmented regions to correct disparities, then homogeneously-colored regions are fixed, but regions with color variations are restricted to planar disparities
Solution Approach 1:
The algorithm segments the image into local neighborhoods based on color homogeneity, allowing different processing for each segment. This enables accurate correction in homogeneous regions while preserving complex surface variations at segment boundaries, avoiding the over-simplification that occurs when forcing entire regions into planar models.
Solution Approach 2:
The algorithm uses iterative refinement where segmentation and disparity correction are dynamically adjusted across multiple passes. Rather than applying a static planar model, the system adapts its processing to the local characteristics of each region in each iteration, allowing complex surfaces to be represented through multiple local corrections rather than a single global plane.
Data Source
AI summary
Methods and apparatus for disparity map correction through statistical analysis on local neighborhoods. A disparity map correction technique may be used to correct mistakes in a disparity or depth map. The disparity map correction technique may detect and mark invalid pixel pairs in a disparity map, segment the image, and perform a statistical analysis of the disparities in each segment to identify outliers. The invalid and outlier pixels may then be corrected using other disparity values in the local neighborhood. Multiple iterations of the disparity map correction technique may be performed to further improve the output disparity map.