Many models of statistical signals are related to the "Markov Model". By this model there is a dependency between a value of specific sample of a signal.

Many models of statistical signals are related to the “Markov Model”. By this model there is a dependency between a value of specific sample of a signal and previous samples of the same signal which are near the current sample. The Markov Model is suited to describe an Image model and connections between pixels’ values in picture, because of images’ nature of local dependency and spatial continuity. Image processing is affected by the nature of the image, therefore, defining a model for the image should improve interpolation of the image comparing to existing interpolation methods.

In this project we will develop an interpolation method of textures. A texture is a structure with elements which are similar to each other but no element is prominent. The texture can be described as a superposition of 2 homogenous, separated, orthogonal components:

- Structural-Global Component – Deterministic component
- Pure Random Component – Pure non-deterministic component

The deterministic component can be described as a superposition of 2 components as well,

- Harmonic Field – generates the periodic features
- Generalized Evanescent Field – generates directional feature

In the project we assume:

__Method A:__

Using a deterministic connection between the Saturation Channel and the Intensity Channel:

- Linear Interpolation of H – Channel
- Texture Interpolation of I – Channel
- Deterministic connection to find S – Channel from Interpolated I – Channel.

**Method A – Block diagram**

**Method A – Results**

__Method B:__

- Converting Color texture to gray-scale texture.
- Texture Interpolation of the gray-scale image.
- Converting back to color texture, using information from original, low-resolution, colored texture.

**Method B – Block diagram**

**Method B – Results**