Image Processing Tone Mapping Using Locally Invariant Space
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current tone-mapping algorithms, such as multi-layer decomposition-based and grid-based methods, face challenges in hardware implementation due to high line-buffer costs and bandwidth requirements, and can produce halos in high contrast edges, making them unsuitable for real-time applications.
Innovation Solution
A method and device for image processing that applies a transfer function to map pixels into a locally invariant image space, using a grid-based local adaptation parameter approach, reducing the need for multiple tone-mapping functions and minimizing hardware requirements by using a single tone-mapping function for all pixels, and employing an inverse transfer function to remap the image back to the original space.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If multi-layer decomposition-based tone-mapping algorithms are used, then image quality is improved, but hardware cost and bandwidth requirements increase
Solution Approach 1:
The patent applies segmentation by dividing the image processing into distinct stages: a transfer function stage that maps pixels to a locally invariant image space, followed by a single tone-mapping function stage, and finally an inverse transfer function stage. This segmentation allows each stage to be optimized independently, reducing overall hardware complexity while maintaining image quality.
Solution Approach 2:
The patent changes the parameter space by introducing a locally invariant image space through the transfer function. By transforming pixels from the original image space to this intermediate space where local contrast properties are normalized, the system can apply a single universal tone-mapping function instead of multiple location-specific functions, thereby reducing hardware requirements.
2Manufacturing precision
If grid-based local adaptation parameter methods are used, then local contrast enhancement is improved, but bandwidth requirements increase
Solution Approach 1:
The patent achieves universality by designing a single tone-mapping function that operates effectively across the entire image after the transfer function has normalized local variations. This single function serves multiple purposes (replacing what would otherwise require multiple location-specific functions), thereby reducing bandwidth requirements for parameter transmission while maintaining local contrast enhancement quality.
3Reliability
If traditional tone-mapping algorithms are used, then image processing capability is maintained, but halos appear in high contrast edges
Solution Approach 1:
The patent introduces the locally invariant image space as an intermediary between the original image space and the final output space. This intermediate representation, created by the transfer function, serves as a mediator that separates the local contrast enhancement task from the global tone-mapping operation, preventing halo artifacts from forming at high contrast edges while maintaining overall image processing capability.
Data Source
AI summary
A method and device for image processing, and storage medium are provided. In the method, an initial image I is acquired; a pixel I(i) of the initial image is selected; a transfer function ƒ is applied to the pixel I(i) to acquire a transferred pixel Ĩ(i), and the transfer function ƒ is determined on the basis of at least the pixel information of the selected pixel; a predetermined tone-mapping function is applied to the transferred pixel Ĩ(i) to acquire a tone-mapped transferred pixel Ĩ′(i); an inverse of the transfer function ƒ−1 is applied to the tone-mapped transferred pixel Ĩ′(i) to acquire a tone-mapped final pixel I′(i). Herein, the steps are repeated for a plurality of pixels of the initial image I to generate a final image I′, and for each of the plurality pixels the same tone-mapping function is applied.