Adaptive Image Filtering for Block Noise Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image decoding technologies face challenges in efficiently coding and decoding images due to block noise at boundaries, which affects coding efficiency and image quality, as they cannot adaptively modify filter coefficients and consider neighboring pixel values effectively.
Innovation Solution
An image decoding apparatus that decodes prediction residual signals and prediction signals, using filter parameters to perform adaptive filtering based on calculated gradient values and weight coefficients, which takes into account the magnitude of transform coefficients and pixel differences across block boundaries.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the number of taps and coefficients of the filter is increased to remove block distortion, then the block noise removal effect is improved, but the image becomes blurred
Solution Approach 1:
The patent applies dynamics by making the filter coefficients adaptive rather than fixed. The filtering apparatus calculates gradient values at block boundaries and selects different filter coefficients from a predetermined set based on the gradient magnitude, allowing the filter strength to dynamically adjust between strong filtering (high gradient) and weak filtering (low gradient), thus resolving the contradiction between block noise removal and image clarity
Solution Approach 2:
The patent changes the parameter of filter coefficients based on gradient values. By selecting different coefficients from a predetermined set according to the calculated gradient magnitude, the system optimizes the balance between blocking effect removal and image quality preservation, transforming a static filter into an adaptive one that responds to local image characteristics
2Reliability
If filtering is applied to remove block noise at boundaries, then block distortion is reduced, but coding efficiency decreases due to information loss during quantization
Solution Approach 1:
The patent applies preliminary action by performing filtering before the blocking effect becomes severe. The filtering apparatus processes block boundaries in the decoded image to remove distortion early in the decoding process, preventing the need for stronger (more blur-inducing) filtering later, thus preserving coding efficiency while reducing block distortion
Solution Approach 2:
The patent uses feedback by calculating gradient values from the decoded image to determine which filter coefficients to apply. This feedback mechanism allows the system to adapt the filtering strength based on the actual block distortion present, optimizing the balance between distortion removal and coding efficiency
3Object-generated harmful factors
If a low-pass filter is used to remove block distortion, then the blocking effect is reduced, but the image becomes blurred conversely
Solution Approach 1:
The patent applies local quality by treating different regions of the image differently based on their gradient characteristics. At block boundaries with high gradients, strong filtering is applied to remove blocking effects. In regions with low gradients (smooth areas), weak or no filtering is applied to preserve sharpness. This localized approach resolves the contradiction between blocking effect removal and image sharpness
Data Source
Figure 1
Figure 2
Figure 3
AI summary
When a filtering process on decoded image data is performed using filter parameters, gradient values near the pixel to be the target of the filtering process are calculated. In accordance with the calculated gradient, a single filter parameter is selected from a set of filter parameters. Based on the selected filter parameter, weight coefficients for the filtering process are calculated. Then, using the pixel value to be the processing target, the surrounding pixel values and the calculated weight coefficients, the pixel value after correction of the processing target pixel is calculated. With this process, it is possible to provide an image coding apparatus that can efficiently code an image properly and an image decoding apparatus that can decode an image, by performing a filtering process in accordance with the image characteristics.