Bitonal Image Compression via Tangent Constraint Matching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional data compression methods for bitonal images struggle to balance compression rate and distortion, particularly in incorporating human perception models effectively, leading to inefficiencies in lossless text compression and image compression.
Innovation Solution
A system and method that compresses bitonal image documents by representing them as sets of components rather than pixels, using human perception models to group similar components and apply context-sensitive matching, thereby reducing the number of models and improving shape matching and segmentation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If conventional pixel-based compression methods are used, then compression can be achieved, but the compression rate is insufficient and distortion is high
Solution Approach 1:
The patent segments the bitonal image into connected components (glyphs, characters, or groups of pixels) rather than processing individual pixels. This segmentation allows for more efficient representation by grouping spatially related pixels into meaningful units that can be compressed more effectively, thereby improving compression rate while maintaining visual fidelity and reducing distortion.
2Reliability
If tighter matchers are applied to reduce mismatches, then matching accuracy improves, but the number of font models increases reducing compression efficiency
Solution Approach 1:
The patent merges similar connected components into shared font models by applying matchers that compare geometric and topological properties. When components are determined to be similar (within acceptable tolerance), they are merged into the same font model, reducing the total number of models needed. This approach maintains matching accuracy by using appropriate matcher strictness while improving compression efficiency through model consolidation.
3Productivity
If looser matchers are applied to reduce the number of models, then compression efficiency improves, but the likelihood of mismatch increases
Solution Approach 1:
The patent employs dynamic matcher selection and adaptive tolerance levels based on the specific characteristics of the connected components being compared. The system adjusts matcher strictness dynamically - using tighter criteria for small, thin components where precision is critical, and looser criteria for large, thick components where slight variations are less visually significant. This dynamic approach optimizes the balance between compression efficiency and matching accuracy.
4Productivity
If human perception models are incorporated, then compression efficiency improves, but computational complexity increases
Solution Approach 1:
The patent applies human perception models selectively to different aspects of the compression process rather than uniformly across all operations. Specifically, perceptual weighting is applied during the matching and distortion assessment phases, where human visual system characteristics (such as sensitivity to certain spatial frequencies or patterns) are used to guide which differences are most important to preserve. This localized application of perceptual models improves compression efficiency by focusing computational resources on perceptually significant features while avoiding unnecessary complexity in less critical areas.
Data Source
AI summary
A system and method of symbol matching may include a processor configured to determine which pixels of a first symbol are tangent pixels; for each of the determined tangent pixels, determine whether a second symbol includes a pixel corresponding to the tangent pixel that includes at least one same tangent constraint as that of the tangent pixel; accept the first and second symbols as a match based on shared tangent constraints conditional upon a determination that the second symbol includes for each of at least a subset of the tangent pixels of the first symbol a corresponding pixel that includes the at least one same tangent constraint as that of the tangent pixel; and generate a document including a single symbol that is mapped to the first and second symbols.