Visual Content-Sensitive Image Encoding for Remote Desktops
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Remote desktop systems face bandwidth limitations and performance issues due to large digital image sizes, which can result in blurred text when compressed using existing image encoding methods, as they often compress images uniformly without distinguishing between text and graphics regions.
Innovation Solution
A visual content-sensitive encoding technique that classifies image blocks as text-oriented or graphics-oriented, using frequency domain parameters to determine appropriate encoding settings, ensuring sharp text while achieving higher compression ratios for graphics regions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of energy
If uniform image compression is applied to reduce bandwidth usage, then network bandwidth consumption is reduced, but text sharpness deteriorates and becomes blurred
Solution Approach 1:
The patent divides the image into multiple blocks and classifies each block as either text-oriented or graphics-oriented. This segmentation allows different compression strategies to be applied to different regions, preventing uniform compression from blurring text while still achieving overall bandwidth reduction through graphics compression.
Solution Approach 2:
The patent applies different encoding quality levels to different image regions based on their content type. Text-oriented blocks receive higher quality encoding to maintain sharpness, while graphics-oriented blocks use higher compression ratios. This local quality differentiation resolves the contradiction between overall bandwidth reduction and text sharpness preservation.
2Productivity
If high compression ratio is applied to reduce image size, then bandwidth efficiency is improved, but image quality deteriorates
Solution Approach 1:
The patent segments the image into text and graphics regions, allowing high compression ratios to be applied specifically to graphics regions while maintaining higher quality for text regions. This segmentation enables bandwidth efficiency improvement without sacrificing overall image quality.
Solution Approach 2:
The patent dynamically changes encoding parameters based on block classification. Text-oriented blocks use encoding parameters optimized for sharpness and readability, while graphics-oriented blocks use parameters optimized for compression ratio. This parameter adaptation resolves the contradiction between bandwidth efficiency and image quality.
3Manufacturing precision
If text-oriented encoding is applied to all image regions, then text sharpness is maintained, but bandwidth consumption increases
Solution Approach 1:
The patent identifies and segments text regions from graphics regions, allowing text-oriented encoding to be applied only where needed rather than uniformly across the entire image. This selective application maintains text sharpness while reducing overall bandwidth consumption by compressing graphics regions more aggressively.
Solution Approach 2:
The patent applies high-quality text-oriented encoding locally to text-oriented blocks while using more compressed graphics-oriented encoding for non-text regions. This local quality differentiation maintains text sharpness where required while optimizing bandwidth usage overall, resolving the contradiction between text sharpness and bandwidth consumption.
Data Source
AI summary
An example method may include identifying a first block of a first image, the first block comprising a plurality of pixel values, generating a frequency-based representation of the first block, where the frequency-based representation comprises a transformation matrix having a plurality of coefficients, where each coefficient specifies a weight of a respective frequency in the frequency-based representation of the first block, generating at least one frequency domain parameter of the first block in view of a sum of a plurality of the coefficients of the transformation matrix, generating a visual content classification value in view of the at least one frequency domain parameter of the first block, selecting, in view of a determination of whether the visual content classification value satisfies a visual content-specific threshold, an encoding, and generating, using the selected encoding, an encoded block in view of the first block.