Depth Map Optimization Using Segmentation and Depth Planes
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing depth map generation algorithms in stereoscopic imaging systems face challenges such as edge blur and high computational and memory resource consumption, leading to inaccuracies and inefficiencies in generating accurate depth information for 3D images.
Innovation Solution
A method that divides images into segmentation maps based on pixel value differences and object distance information, partitions the depth map into depth planes, and fills holes with optimized depth values using segmentation maps with varying fineness, resulting in an optimized depth map with reduced noise and improved accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If local-correlation algorithm is used for depth map generation, then calculation amount is reduced, but edge blur issue occurs and depth accuracy deteriorates
Solution Approach 1:
The patent divides the depth map into multiple depth planes based on depth ranges, and further segments each depth plane into multiple regions using segmentation maps. This multi-level segmentation allows different processing strategies to be applied to different regions, improving depth accuracy without requiring full-image complex processing, thus balancing calculation amount and precision.
Solution Approach 2:
The patent applies different processing methods to different regions within depth planes. Specifically, boundary regions are identified and processed differently from interior regions, with boundary regions receiving more sophisticated hole-filling operations. This local quality approach improves depth accuracy at critical edges while maintaining efficiency in interior regions.
2Measurement precision
If pixel-wise matching method is used for depth map generation, then depth accuracy is improved, but calculation amount and memory resource consumption increase significantly
Solution Approach 1:
The patent segments the image into multiple depth planes and further into regions, allowing pixel-wise matching to be applied selectively only to boundary regions where high accuracy is critical, rather than the entire image. This reduces overall calculation amount while maintaining depth accuracy where most needed.
Solution Approach 2:
The patent applies high-precision pixel-wise matching methods locally to boundary regions identified through segmentation, while using simpler methods for interior regions. This local quality strategy improves depth accuracy at edges without incurring the full computational cost of applying pixel-wise matching globally.
3Measurement precision
If segmentation maps with fine dividing fineness are used, then depth accuracy in boundary regions is improved, but calculation amount and processing complexity increase
Solution Approach 1:
The patent implements a two-level segmentation system: first dividing the depth map into depth planes, then further segmenting each plane into regions. This hierarchical segmentation manages complexity by breaking down the problem into smaller, more manageable parts that can be processed independently with appropriate complexity levels.
Solution Approach 2:
The patent applies fine dividing fineness segmentation specifically to boundary regions where high precision is needed, while using coarser segmentation for interior regions. This local quality approach improves depth accuracy at boundaries without incurring the full processing complexity of fine segmentation across the entire image.
4Ease of operation
If depth map is processed as a whole, then processing is simple, but noise reduction and accuracy improvement are limited
Solution Approach 1:
The patent divides the depth map into multiple depth planes based on depth ranges, and further segments each plane into regions. This segmentation enables different processing operations to be applied to different parts of the depth map, improving noise reduction and accuracy while maintaining manageable processing complexity through systematic organization.
Solution Approach 2:
The patent applies different processing strategies to different regions within depth planes. Boundary regions receive specialized hole-filling and noise reduction processing, while interior regions are processed with simpler methods. This local quality approach improves overall depth accuracy without requiring complex processing of the entire depth map uniformly.
Data Source
AI summary
Method and apparatus for optimizing depth information are provided. One of a left image and a right image is divided into a plurality of segmentations for obtaining a plurality of segmentation maps. A necessary repair depth map is obtained, and the necessary repair depth map is partitioned into a plurality of depth planes according to a plurality of primary depth values and a camera parameter. The primary depth values are recorded in the necessary repair depth map having a plurality of holes. A plurality of optimized depth values are respectively generated for the holes in each of the depth planes by using the segmentation maps, and the optimized depth values are filled into the depth planes to obtain an optimized depth map.