Global Color Image Enhancement Algorithm Based on Retinex Model

Retinex as a luminance perceptual model based on color consistency is introduced. Illumination image is evaluated using a fast Gaussian filtering and global γ transform to original image value of HSV space. Then the reflection image including intrinsic properties of objects is obtained and image enhancement is achieved. Visual observation and quantitative analysis show that the global enhancement algorithm has significant effect on images affected by real illumination factor.