Depth Map Coding Distortion Prediction for Free Viewpoint Video
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for predicting depth map coding distortion in two-dimensional free viewpoint videos are inefficient, leading to high calculation complexity and limited accuracy, especially in real-time systems, as they do not effectively consider the impact on synthesized views and are based on one-dimensional models.
Innovation Solution
A method that synthesizes texture maps from multiple viewpoints to generate distortion prediction weights, allowing for accurate estimation of depth map coding distortion by averaging distortion sums in different directions, thereby reducing calculation complexity and improving prediction accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If conventional distortion calculation methods (SSD, SAD) are used, then implementation simplicity is improved, but prediction accuracy deteriorates because they do not consider the impact on synthesized views
Solution Approach 1:
The patent introduces an intermediary synthesis algorithm that simulates the view synthesis process to calculate distortion. This intermediary step bridges the gap between simple distortion metrics and accurate synthesized view quality assessment, allowing the system to achieve both computational feasibility and high prediction accuracy by using the synthesis algorithm as a mediator to evaluate distortion impact.
Solution Approach 2:
The patent performs preliminary view synthesis before distortion calculation to pre-compute the synthesized view from reference viewpoints. This preliminary action enables the system to prepare the necessary synthetic image data in advance, which is then used to calculate distortion metrics accurately without requiring repeated synthesis operations during the actual distortion measurement process.
2Measurement precision
If view synthesis algorithm is executed repeatedly to calculate depth map distortion, then prediction accuracy is improved, but calculation complexity deteriorates significantly
Solution Approach 1:
The patent executes the view synthesis algorithm in advance to generate synthesized views from reference viewpoints before the actual distortion calculation. By performing this synthesis operation as a preliminary action, the system pre-computes the necessary synthetic images that will be used for distortion measurement, thereby avoiding the need to repeatedly execute the complex synthesis algorithm during multiple distortion calculations.
Solution Approach 2:
The patent creates a copy of the synthesized view from reference viewpoints and uses this copied synthetic image for distortion calculation. Instead of repeatedly executing the full synthesis algorithm, the system uses the pre-generated synthetic view as a reference copy to measure distortion, significantly reducing the computational burden while maintaining accurate distortion prediction.
3Device complexity
If one-dimensional free viewpoint video models are used, then model simplicity is improved, but adaptability deteriorates for two-dimensional free viewpoint videos
Solution Approach 1:
The patent extends the distortion calculation model from one-dimensional to two-dimensional by incorporating synthesized views in multiple directions. Instead of only considering distortion along a single dimension, the system now calculates distortion in both horizontal and vertical dimensions by synthesizing views from multiple reference viewpoints, thereby adapting the model to handle two-dimensional free viewpoint videos while maintaining the underlying simplicity of the distortion measurement approach.
Data Source
AI summary
Disclosed is a method for predicting depth map coding distortion of a two-dimensional free viewpoint video, including: inputting sequences of texture maps and depth maps of two or more viewpoint stereoscopic videos; synthesizing a texture map of a first intermediate viewpoint of a current to-be-coded viewpoint and a first adjacent viewpoint, and synthesizing a texture map of a second intermediate viewpoint of the current to-be-coded viewpoint and a second adjacent viewpoint by using a view synthesis algorithm; recording a synthetic characteristic of each pixel according to the texture map and generating a distortion prediction weight; and calculating to obtain total distortion according to the synthetic characteristic and the distortion prediction weight.