AI-Weighted Image Combining for Rendering Noise Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image rendering methods, such as Monte Carlo path tracing, suffer from noise and require numerous samples, leading to long processing times and inefficiencies in noise reduction, particularly when attempting to remove residual noise and system errors in correlated images.
Innovation Solution
An apparatus and method that combines independent and correlated images using artificial intelligence to determine pixel-specific weights for enhancing image quality, leveraging an AI apparatus to train and combine images based on normal, texture, and depth information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If Monte Carlo path tracing method is used for realistic rendering, then image realism is improved, but noise is introduced and processing time increases
Solution Approach 1:
The patent segments the image processing into two distinct components: independent image generation (providing realism) and correlated image generation (providing noise reduction). By dividing the rendering process into these separate stages with different sampling strategies, the system achieves both realism and efficiency without requiring excessive samples in a single pass.
Solution Approach 2:
The patent merges independent image data and correlated image data through a combination process that integrates results from two different rendering approaches. This merging allows the system to leverage the realism strengths of independent sampling while incorporating the noise-reduction benefits of correlated sampling, achieving both goals simultaneously.
2Measurement precision
If many samples are used in Monte Carlo path tracing, then image quality is improved, but processing time increases significantly
Solution Approach 1:
The patent applies partial action by using a moderate number of samples for independent rendering and a different sampling strategy for correlated rendering. Rather than using excessive samples in a single pass, the system distributes sampling efforts across two passes, each optimized for its specific purpose, achieving high quality without excessive computational cost in either pass.
Solution Approach 2:
The patent maintains continuity of useful action by performing independent rendering and correlated rendering as complementary processes that both contribute to the final result. Rather than performing one extensive rendering pass, the system continuously produces useful intermediate results from both rendering approaches that are then combined, maximizing productivity throughout the process.
3Object-affected harmful factors
If correlated image processing is used to reduce noise, then noise reduction is improved, but residual noise and system errors persist
Solution Approach 1:
The patent introduces an intermediary combination process that mediates between independent image data and correlated image data. This intermediary step allows the system to process both types of data through a unified framework that can address the limitations of each individual approach, reducing residual noise and system errors that would persist if only correlated processing were used.
Solution Approach 2:
The patent creates a composite rendering approach by combining results from independent rendering and correlated rendering. Similar to composite materials in physics, this composite approach integrates the strengths of both methods while mitigating their individual weaknesses, producing a final image that has reduced noise without the residual errors characteristic of single-method approaches.
4Productivity
If independent image rendering is used, then processing speed is improved, but noise and lack of pixel correlation reduce image quality
Solution Approach 1:
The patent segments the rendering task into independent rendering (for speed) and correlated rendering (for quality), allowing each segment to be optimized for its specific strength. The independent rendering pass processes quickly with minimal correlation assumptions, while the correlated rendering pass focuses on quality enhancement, and the results are combined to achieve both speed and quality.
Data Source
AI summary
An apparatus for enhancing image quality includes a combining module configured to combine an independent image having pixels independent from each other and a correlated image including correlation information between pixels and an artificial intelligence (AI) apparatus configured to provide a weight used to combine the independent image and the correlated image.