Exposure-Bracketed Image Blending via Weight Distribution Functions
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for managing and blending exposure-bracketing images are computationally expensive, complex, and unsuitable for non-professional equipment, often requiring specialized hardware and being cumbersome to use.
Innovation Solution
A system and method for blending digital input images of a scene by selecting corresponding pixels from multiple images, assigning weights based on image characteristics using a weight distribution function, and modifying pixel values to create a new image with a fuller dynamic range, allowing for user-controlled adjustments through a user interface.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If current HDR techniques are used to capture full dynamic range, then image quality is improved, but memory and processing demands increase significantly
Solution Approach 1:
The patent transforms HDR processing by changing the parameter representation from high-bit-depth floating-point values to standard 8-bit per channel pixel values. This parameter transformation maintains the full dynamic range information while reducing memory and processing requirements to levels compatible with consumer devices.
Solution Approach 2:
The patent creates a simplified copy of the HDR blending process that operates on standard 8-bit images rather than requiring full HDR data structures. The weight distribution functions and blending algorithms are adapted to work with conventional image formats, producing HDR-quality results without HDR-specific hardware or software requirements.
2Manufacturing precision
If complex HDR algorithms are applied, then blending accuracy is improved, but computational cost increases
Solution Approach 1:
The patent extracts the essential blending function from complex HDR algorithms by isolating the weight assignment mechanism. Instead of implementing full HDR processing pipelines, the invention uses weight distribution functions applied to standard 8-bit images, removing unnecessary computational complexity while preserving blending accuracy.
Solution Approach 2:
The patent replaces expensive, complex HDR data structures and processing routines with simple 8-bit image manipulations. The weight distribution functions use basic arithmetic operations on standard image pixels, making the process computationally inexpensive and suitable for real-time or near-real-time processing on consumer hardware.
3Measurement precision
If specialized HDR workflows are implemented, then image processing quality is improved, but ease of operation deteriorates
Solution Approach 1:
The patent makes HDR-style image blending universally accessible by implementing it within existing consumer software platforms (Adobe Photoshop, GIMP, etc.). The same blending algorithms work across different image formats and software applications, eliminating the need for specialized HDR workflows or dedicated software tools.
Solution Approach 2:
The patent enables automated HDR blending through weight distribution functions that operate automatically on selected images without requiring manual HDR-specific processing steps. Users simply select multiple exposure images, and the system automatically applies the blending algorithm, removing the complexity of manual HDR workflow management.
4Measurement precision
If many bits per color channel are used, then color accuracy is improved, but memory requirements increase
Solution Approach 1:
The patent changes the bit-depth parameter from 16 or 32 bits per channel (required for full HDR) to 8 bits per channel (standard consumer format). This parameter change is achieved through the weight distribution function methodology, which preserves color accuracy information through weighted averaging while storing results in compact 8-bit channels, reducing memory usage by 75-87%.
Data Source
AI summary
Systems and methods are presented for generating a new digital output image by blending a plurality of digital input images capturing the same scene at different levels of exposure. Each new pixel for the new digital output image is derived from a group of corresponding aligned pixels from the digital input images. In order to determine a weight for each pixel in each group of mutually-aligned source-image pixels, a weight distribution function is applied to values of an image characteristic for the pixels in the group of corresponding aligned pixels, and a net weight is subsequently assigned to each of the pixels in the group. Pixel values of pixels in each group of mutually-aligned source-image pixels are modified based on the net weights assigned to the pixels in order to obtain a new pixel value for a corresponding new pixel in the new digital output image.