BREN: body reflection essence-neuter model for separation of reflection components

Area of Science:

• Computer Vision
• Image Processing
• Color Science

Background:

• Separating reflection components in images is crucial for accurate color analysis.
• Existing methods often struggle with noise or specific color spaces like RGB and CMY (cyan, magenta, and yellow).

Purpose of the Study:

• To propose a novel reflection color model and an effective method for separating dichromatic reflection components from a single image.
• To develop a method that is robust to noise and applicable to a wide range of colors.

Main Methods:

• Introduced a reflection color model based on 'body essence' and 'neuter' components.
• Developed a method utilizing a Gaussian function to separate interface reflection, avoiding thresholding issues.
• The method operates on a single image without requiring region segmentation.

Main Results:

• The proposed model and method effectively separate reflection components.
• The approach demonstrates insensitivity to noise.
• It is suitable for colors in both RGB and CMY spaces, outperforming maximum chromaticity-based methods.

Conclusions:

• The novel reflection color model and separation method are effective and robust.
• This technique offers an improvement over existing methods for dichromatic reflection separation.
• The method's applicability across color spaces and noise resilience make it a valuable tool in image processing.