QRetinex-Net: Quaternion-Valued Retinex Decomposition for Low-Level Computer Vision Applications
This work addresses low-level computer vision challenges for applications such as inspection and detection, but it is incremental as it builds on existing Retinex theory with a novel mathematical formulation.
The paper tackled the problem of low-light image degradation by proposing a quaternion-valued Retinex decomposition to address flaws in classic models, resulting in gains of 2-11% over leading methods in applications like crack inspection and face detection with improved color fidelity and noise reduction.
Images taken in low light often show color shift, low contrast, noise, and other artifacts that hurt computer-vision accuracy. Retinex theory addresses this by viewing an image S as the pixel-wise product of reflectance R and illumination I, mirroring the way people perceive stable object colors under changing light. The decomposition is ill-posed, and classic Retinex models have four key flaws: (i) they treat the red, green, and blue channels independently; (ii) they lack a neuroscientific model of color vision; (iii) they cannot perfectly rebuild the input image; and (iv) they do not explain human color constancy. We introduce the first Quaternion Retinex formulation, in which the scene is written as the Hamilton product of quaternion-valued reflectance and illumination. To gauge how well reflectance stays invariant, we propose the Reflectance Consistency Index. Tests on low-light crack inspection, face detection under varied lighting, and infrared-visible fusion show gains of 2-11 percent over leading methods, with better color fidelity, lower noise, and higher reflectance stability.