Integer Image Scaler Reducing Floating-Point Computational Intensity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image scaling methods require computationally intensive floating-point operations, which can be resource-heavy for general-purpose processors, and often involve scaling each pixel independently, increasing computational intensity further.
Innovation Solution
A compiled image scaler using integer and bitwise operations, employing quad-linear or bi-linear filtering to blend colors of adjacent pixels, which reduces the need for floating-point operations and allows concurrent scaling of color channels, thereby decreasing computational intensity by leveraging integer registers and preprocessor macros.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If floating-point operations are used for image scaling, then image quality is improved, but computational intensity and resource consumption increase
Solution Approach 1:
The patent changes the numerical parameter type from floating-point to fixed-point representation. This allows the system to maintain image scaling functionality while using simpler integer arithmetic operations instead of complex floating-point operations, thereby reducing computational intensity while preserving acceptable image quality through careful selection of fixed-point precision levels.
Solution Approach 2:
The patent substitutes floating-point arithmetic operations with integer arithmetic operations. By replacing the floating-point computational mechanism with integer-based fixed-point arithmetic, the system achieves the same image scaling function with significantly reduced computational complexity and resource consumption, as integer operations are natively supported by processor architectures without requiring specialized floating-point units.
2Ease of operation
If each pixel is scaled independently, then processing simplicity is improved, but computational intensity increases
Solution Approach 1:
The patent merges the processing of multiple pixels by exploiting spatial redundancy. Instead of treating each pixel independently, the method combines calculations for adjacent pixels by sharing common intermediate results and using overlapping computation windows, thereby reducing the total number of operations while maintaining processing simplicity through unified algorithmic treatment.
Solution Approach 2:
The patent performs preliminary computation of shared values that will be used by multiple pixels. By pre-calculating intermediate results that are common to several pixel operations and storing them for reuse, the system reduces redundant computations and lowers overall computational intensity while keeping the processing approach systematic and manageable.
3Adaptability or versatility
If conventional image scaling methods are used, then compatibility is improved, but processing speed decreases
Solution Approach 1:
The patent substitutes conventional floating-point scaling operations with fixed-point integer operations. This substitution maintains compatibility with standard processor architectures that natively support integer arithmetic, while achieving faster processing speeds because integer operations can be executed more quickly and with lower overhead than floating-point operations on general-purpose processors.
Solution Approach 2:
The patent changes the numerical representation parameter from floating-point to fixed-point format. This parameter change enables the use of optimized integer arithmetic instructions that are faster and more efficient on typical hardware, thereby increasing processing speed while maintaining sufficient precision for image scaling applications through appropriate fixed-point bit allocation.
Data Source
AI summary
In the described examples, a compiled image scaler includes a set of machine executable instructions that generate a scaled image that is a scaled version of a source image with integer and bitwise operations. The compiled image scaler employs filtering to blend colors of adjacent pixels in the source image to generate the scaled image, and each filtering operation concurrently scales three color channels of a pixel in the source image.