Adaptive Image Rendering Using a Nonlinear Mapping-Function-Based Retinex Model

Area of Science:

• Computer Vision
• Image Processing
• Computational Photography

Background:

• Low-light image enhancement is crucial for various applications.
• Existing retinex-based methods often suffer from halo artifacts and limited contrast control.
• Adaptive rendering techniques are needed to address varying illumination conditions.

Purpose of the Study:

• To introduce an adaptive image rendering method for low-light enhancement.
• To reduce halo artifacts commonly found in retinex-based approaches.
• To adaptively control image contrast for improved visual quality.

Main Methods:

• Utilized a parametric nonlinear mapping function based on the retinex model.
• Estimated reflectance component using only the luminance channel to minimize artifacts.
• Proposed a novel nonlinear mapping function incorporating luminance and reflectance statistics.
• Developed a method for adaptive gain and offset determination to control contrast.
• Reconstructed chrominance channels based on the relationship between estimated and reconstructed luminance.

Main Results:

• Significantly reduced halo artifacts compared to traditional methods.
• Achieved adaptive contrast control for enhanced image appearance.
• Demonstrated subjective and objective improvements in low-light image rendering.
• Outperformed state-of-the-art scale-based retinex methods in experimental evaluations.

Conclusions:

• The proposed adaptive retinex-based method effectively enhances low-light images.
• The technique offers superior artifact reduction and contrast control.
• Results indicate a promising advancement in low-light image processing.