Guided Image Filtering via Down-sampling and Resolution Parameters
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing guided filter technologies require high computational complexity and cost due to the need for multiple neural networks, increasing implementation difficulty.
Innovation Solution
An image processing method and device that down-samples an image, obtains resolution parameter combinations, and converts target pixels into output pixels using selected candidate pixels, reducing the need for multiple neural networks and simplifying the processing steps.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If multiple neural networks are used to implement guided filter, then filtering performance is improved, but computational complexity and cost increase
Solution Approach 1:
The patent extracts and removes the complex neural network components from the guided filter implementation. Instead of using multiple neural networks to achieve guided filtering, the invention uses a simplified approach with a single neural network that processes only the guide image, eliminating the need for multiple network passes and reducing overall computational complexity while maintaining filtering effectiveness
Solution Approach 2:
The patent replaces expensive and complex neural network models with a simpler, more efficient processing approach. The invention uses a single neural network model that is computationally less intensive, effectively substituting the expensive multiple-network solution with a cheaper, more efficient alternative that achieves similar filtering results
2Measurement precision
If multiple neural networks are used for guided filter, then filtering accuracy is improved, but implementation difficulty increases
Solution Approach 1:
The patent extracts and eliminates the complex multi-network architecture from the implementation. By removing the requirement for multiple neural networks and using a single network that processes only the guide image, the invention significantly simplifies the implementation process while maintaining the accuracy needed for effective guided filtering
Solution Approach 2:
The patent creates a universal processing framework where a single neural network performs multiple functions that previously required separate networks. The single network handles the guide image processing and parameter generation, eliminating the need for multiple specialized networks and reducing implementation complexity
Data Source
AI summary
Embodiments of the disclosure provide an image processing method and an image processing device. The method includes the following. A first image is obtained, and the first image is down-sampled into a second image. A first resolution parameter combination of each second pixel in the second image is obtained. A target pixel among the first pixels is selected, and second pixels corresponding to the target pixel among the second pixels are accordingly obtained. A comparison result is determined between the target pixel and each second pixel corresponding to the target pixel, and a candidate pixel among the second pixels corresponding to the target pixel is accordingly selected. A second resolution parameter combination of the target pixel is selected based on the first resolution parameter combination of the candidate pixel. The target pixel is converted into an output pixel based on the second resolution parameter combination.