Milano-Retinexes are spatial color algorithms based on the original Retinex model. They enhance any input image by processing the local, relative channel intensity differences and pixel spatial positions. The Milano-Retinex approach presented here, catches these local features by exploring the image with paths, which locally minimize an energy functional. The contributions of intensity and spatial cues are tunable by the user: this enables to investigate the effect of these features on the overall image enhancement.
This is poster number 57 in Poster Session