Mathematical insights into the original Retinex algorithm for image enhancement

Area of Science:

• Computer Vision
• Image Processing
• Color Science

Background:

• The Retinex theory, a computational model of human color perception, is foundational to digital image enhancement.
• The Retinex algorithm improves image quality by enhancing detail visibility, color, and illumination effects, but lacks a complete mathematical formulation.
• Existing mathematical formalizations offer partial renderings and model variants.

Purpose of the Study:

• To provide a comprehensive mathematical formalization of the original Retinex algorithm.
• To offer mathematical insights into Retinex theory and its application in image enhancement.
• To facilitate comparison between the original Retinex model and later variants.

Main Methods:

• Formalizing the original Retinex algorithm using mathematical principles.
• Comparing the formalized model with existing Retinex-based models in the literature.
• Analyzing differences and similarities with models like Provenzi et al. (2005).

Main Results:

• A complete mathematical framework for the original Retinex algorithm has been established.
• The formalization clarifies the underlying mathematical principles of Retinex.
• Comparative analysis highlights the distinct contributions of the original model versus later adaptations.

Conclusions:

• The complete mathematical formalization enhances understanding of the original Retinex algorithm.
• This work promotes awareness of Retinex's utility in digital image enhancement.
• It provides a basis for appreciating the evolution and variations of Retinex-based image processing models.