Wavelet Image Compression Null Post Distortion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Lossy wavelet-based compression techniques often distort image data containing null posts, textual or graphical overlays, and other features distinct from the underlying image, leading to blurring and aliasing effects due to quantization, which complicates accurate reconstruction of the original image data.
Innovation Solution
A method that identifies and modifies wavelet coefficients associated with null posts and overlays during the forward wavelet transform, replacing them with replacement coefficients to minimize distortion, and uses an inverse wavelet transform to generate modified image data that can be compressed with lossy wavelet-based compression while preserving the underlying image data, allowing for accurate reconstruction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If lossy wavelet-based compression is applied to image data containing null posts or overlays, then compression performance is improved, but distortion and blurring effects occur in the decompressed image
Solution Approach 1:
The patent segments the image data into two distinct categories: valid image data and special data values (null posts, overlays). By identifying and separating these different types of data, the patent applies different processing strategies to each segment, allowing lossy compression for valid data while preserving special values through a placeholder mechanism, thus resolving the contradiction between compression efficiency and reconstruction accuracy
Solution Approach 2:
The patent changes the parameter representation of special data values by replacing extreme values (such as -32768 for null posts) with placeholder values that have specific bit patterns (all zeros or all ones). This parameter transformation allows the compression algorithm to recognize and preserve these special values during lossy compression, preventing distortion while maintaining compression performance
2Manufacturing precision
If lossless compression is applied to regions of interest containing null posts or overlays, then image distortion is reduced, but file size increases
Solution Approach 1:
The patent applies local quality by treating different regions of the image with different compression strategies. Valid image regions undergo lossy wavelet compression, while regions containing special data values are handled through placeholder replacement. This localized approach ensures high fidelity where needed (preserving special values) while maintaining overall compression efficiency for the majority of valid image data
Data Source
AI summary
Techniques are provided for compressing and decompressing image data which may reduce the distortion that may otherwise be created by the compression of data values representative of null posts, overlays or other features that differ from the underlying image. In compression, at least one coefficient generated by a forward wavelet transform may be replaced with respective replacement coefficients. The transformed image data is then subjected to an inverse wavelet transform to generate modified image data in which the data values which differ from the underlying image have been replaced by interpolated or extrapolated values. The modified image data may be compressed in accordance with wavelet-based image compression. Alternatively, wavelet image compression may be applied directly to the coefficients resulting from the modified forward wavelet transform. In decompression, the compressed image data may be decompressed and data values representative of null posts or other features may be replaced with their original values.