Progressive Text Encoding for Low-Latency Image Transmission
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Transmission of high-resolution images with text information over constrained channels faces challenges such as latency and inefficiency, particularly in real-time interactive systems, as existing methods like JPEG are not suitable for preserving text detail and often result in noticeable delays.
Innovation Solution
A method that analyzes image pixels to identify text regions, generates masks to differentiate between text and background pixels, and uses progressive encoding techniques to transmit text information efficiently, allowing for early display of text regions with refinement updates, thereby reducing latency and improving responsiveness.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If lossy compression methods like JPEG are used to reduce data size, then transmission efficiency is improved, but text detail preservation deteriorates
Solution Approach 1:
The patent segments the image into distinct text regions and non-text regions. Text regions are identified through analysis of pixel characteristics such as edge density, color uniformity, and spatial patterns. This segmentation allows different compression strategies to be applied to different regions, preserving text quality while improving overall transmission efficiency.
Solution Approach 2:
The patent applies different compression quality levels to different regions of the image. Text regions are encoded with lossless or high-quality lossy compression to preserve character details, while non-text regions use more aggressive lossy compression. This local quality approach resolves the contradiction by maintaining text precision where needed while improving overall productivity through efficient compression of less critical areas.
2Measurement precision
If high-resolution images are transmitted to preserve detail, then image quality is improved, but transmission latency increases
Solution Approach 1:
The patent transmits only the necessary portions of image data at high resolution. By identifying and prioritizing text regions, the system transmits full-detail data only for areas containing text, while using compressed or lower-resolution data for non-text areas. This partial action approach maintains image quality where needed while reducing overall transmission latency.
Solution Approach 2:
The patent creates simplified representations or copies of non-text regions that require less transmission bandwidth. Instead of transmitting full-resolution data for all areas, the system generates compressed copies or approximations for non-critical regions, reducing transmission time while maintaining acceptable quality for those areas.
3Measurement precision
If lossless encoding is applied to text regions to preserve detail, then text quality is improved, but transmission efficiency deteriorates
Solution Approach 1:
The patent segments the image to identify text regions that require lossless encoding. By precisely delimiting text areas through pixel analysis, the system applies computationally intensive lossless compression only to these specific regions, while using more efficient lossy compression for the remainder of the image, thus balancing text quality with overall transmission efficiency.
Solution Approach 2:
The patent implements local quality differentiation by applying lossless encoding specifically to text regions where precision is critical, while applying lossy encoding to non-text regions where some quality degradation is acceptable. This selective approach optimizes the balance between text quality preservation and transmission efficiency across the entire image.
Data Source
AI summary
A method for encoding an image, one embodiment comprising analyzing colors and spatial features of pixels of the image to identify a text region separate from a picture region; generating, for a portion of the text region, a mask dividing the portion into background pixels and text pixels, the background pixels identified as pixels with a constant color, the text pixels identified as pixels contrasting the constant color and comprising a plurality of colors; analyzing chrominance values of the plurality of colors to determine a text chrominance; generating, for each text pixel of the plurality of text pixels, a text pixel value to generate text pixel values, each text pixel value based on a luminance of a text pixel for which it was generated; and transmitting an encoding of the portion comprising an encoding of each of the constant color, the mask, the text chrominance and the text pixel values.