Adaptive Block Filtering for Video Encoding Noise Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image and video denoising methods struggle to effectively address encoding noise, particularly the blocking effect, and often require significant computational resources, leading to low real-time performance.
Innovation Solution
A method for processing images and videos that involves training a filter model using the Least Squares method to adaptively filter each image block based on its specific encoding noise intensity and attribute, enabling real-time noise reduction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional denoising methods are used to eliminate encoding noise, then noise reduction is achieved, but computational resources are significantly consumed leading to low real-time performance
Solution Approach 1:
The image is divided into multiple image blocks, and each block is processed independently with its own filter model. This segmentation allows parallel processing and reduces the overall computational burden while maintaining effective noise reduction for each region.
Solution Approach 2:
Different filter models are trained and applied to different image blocks based on their specific characteristics (edge, texture, smooth regions). This local adaptation ensures optimal noise reduction for each region without requiring computationally expensive global processing.
2Ease of operation
If highly compressed video streams are used to accommodate poor network conditions, then transmission feasibility is improved, but visible encoding noise appears reducing video quality
Solution Approach 1:
The patent converts the harmful encoding noise into a detectable pattern by training filter models specifically on noisy image blocks. The noise characteristics become the basis for developing targeted filter models that can effectively remove them, turning the degradation into an opportunity for specialized processing.
Solution Approach 2:
The patent adjusts filtering parameters dynamically by training separate filter models for different types of image blocks (edge, texture, smooth regions). This parameter adaptation allows the system to maintain video quality by applying appropriate filtering strength and type to each region, compensating for the compression artifacts.
3Measurement precision
If adaptive filtering is applied to each image block based on its characteristics, then filtering accuracy is improved, but device complexity increases
Solution Approach 1:
The image processing system is segmented into multiple independent filter models, each dedicated to a specific type of image block. This segmentation simplifies the complexity by breaking down the complex adaptive filtering task into multiple simpler, specialized sub-tasks that can be executed in parallel.
Data Source
Figure 1~2
Figure 3~4
Figure 5
AI summary
Disclosed are an image processing method and apparatus, a mobile terminal video processing method and apparatus, a device and a medium. The image processing method comprises: determining an image attribute of each of a plurality of image blocks in an image to be processed; determining, according to the image attribute of each image block, a filtering model corresponding to each image block; and filtering image block by means of the filtering model corresponding to each image block, so as to obtain a filtered image of the image to be processed.