Logarithmic Image Processing for Color Images
| Autor: | Frederic Itthirad, Michel Jourlin, B. Closs, Josselin Breugnot, Mohamed Bouabdellah |
|---|---|
| Rok vydání: | 2011 |
| Předmět: |
Discrete mathematics
Color histogram Pure mathematics Channel (digital image) Color image Computer Science::Computer Vision and Pattern Recognition Scalar (mathematics) ComputingMethodologies_IMAGEPROCESSINGANDCOMPUTERVISION Scalar multiplication Edge detection Vector space Interpolation Mathematics |
| Popis: | The logarithmic image processing (LIP) framework is classically devoted to grey-scale images. This study extends this framework to color images. This new model is denoted LIPC: LIP-color. It does not consist of applying the LIP model to each channel R, G, and B of a color image. We define the transmittance of color images to provide a physical justification, on which we base the definition of logarithmic operators such as addition, subtraction, and scalar multiplication, respectively noted in the LIPC as c , c , and c . As for the classical LIP model, the laws c and c define a vector space structure on the space of images that enables us to present notions requiring such a structure. For example, we define a color logarithmic interpolation by associating with a pair ( F , G ) of images the interval [ F , G ] set of barycenters of F and G . A new notion of color contrast is defined, which satisfies sub-additivity and homogeneity for scalar multiplication. This notion is proved to be efficient for edge detection. We note that the vector space structure opens the way for much new development concerning the definition of metrics, norms, scalar products, and so on and the transfer to LIPC gauges theory, duality theory, and so forth. In this chapter, we focus on applications of the LIPC model. For example, color prediction is presented and discussed, in addition to stabilization of images by dynamic range centering and enhancement of underlighted images. Concerning the implementation of the LIPC operators and algorithms, information is provided on their execution time. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|