Multi-scale color texture image segmentation method combined with MRF (Markov Random Field) and neural network

Abstract

The invention relates to the field of image processing and discloses a multi-scale color texture image segmentation method combined with an MRF (Markov Random Field) and a neural network. The method is mainly used for solving the problem that a simple Gauss model can not be used for accurately describing a feature field distribution characteristic in the traditional MRF method. The method comprises the following steps: when modeling a multi-scale feature field, estimating a probability distribution of a texture feature by using the output of the neural network in a supervision environment on each scale; when modeling a multi-scale mark field based on a classic Potts model, considering the interaction of a mark set of a same-scale second-order neighborhood position and a corresponding mark on a lower resolution scale for each pixel position; by using the maximum likelihood criterion, gradually acting a segmentation result on the lower resolution scale on the next scale from top to bottom to finally realize the multi-scale image segmentation. The segmentation result obtained by the invention has good region homogeneity and Boundary authenticity, and the method can be used for segmenting the color texture image.

Description

technical field

[0001] The invention belongs to the technical field of image processing, and relates to an image segmentation method, which can be used for the segmentation of color texture images. Background technique

[0002] Texture plays a very important role in many applications of computer vision and image processing. Texture segmentation is to divide texture images into different texture regions. It is widely used in many fields, such as remote sensing image information extraction, document image analysis, Restoration of shape information and content-based image retrieval, etc.

[0003] In the past few decades, the method of texture segmentation combined with Markov random field model (MRF) model under the Bayesian framework has received great attention from researchers at home and abroad. Among them, the most classic is the segmentation method based on the double random field model. It uses different feature fields to model the texture features of different regions...

Examples

Embodiment Construction

[0029] Concrete realization process of the present invention is as follows:

[0030] Step 1: Input the image to be segmented, and extract the R, G, and B values ​​of each pixel at a given scale s and the size of the pixel as w s ×w s In the neighborhood of (w s Neighborhood window size) The spectral mean and standard deviation of the three bands of R, G, and B form a feature vector, and its specific execution process is as follows:

[0031] (1a) According to the given scale s, determine the neighborhood size as w s ×w s ;

[0032] (1b) Calculate the spectral mean and standard deviation of the R, G, and B bands in the neighborhood pixel by pixel of the image to be segmented, and the mean value is:

[0033]

[0034] The standard deviation is:

[0035]

[0036] Among them, v ∈ {r, g, b} represents a band of the texture image, ij represents the current pixel position, and w is the neighborhood window diameter;

[0037] (1c) For each pixel position (i, j) of the image,...

technical field

[0001] The invention belongs to the technical field of image processing, and relates to an image segmentation method, which can be used for the segmentation of color texture images. Background technique

[0002] Texture plays a very important role in many applications of computer vision and image processing. Texture segmentation is to divide texture images into different texture regions. It is widely used in many fields, such as remote sensing image information extraction, document image analysis, Restoration of shape information and content-based image retrieval, etc.

[0003] In the past few decades, the method of texture segmentation combined with Markov random field model (MRF) model under the Bayesian framework has received great attention from researchers at home and abroad. Among them, the most classic is the segmentation method based on the double random field model. It uses different feature fields to model the texture features of different regions...

Examples

Embodiment Construction

[0029] Concrete realization process of the present invention is as follows:

[0030] Step 1: Input the image to be segmented, and extract the R, G, and B values ​​of each pixel at a given scale s and the size of the pixel as w s ×w s In the neighborhood of (w s Neighborhood window size) The spectral mean and standard deviation of the three bands of R, G, and B form a feature vector, and its specific execution process is as follows:

[0031] (1a) According to the given scale s, determine the neighborhood size as w s ×w s ;

[0032] (1b) Calculate the spectral mean and standard deviation of the R, G, and B bands in the neighborhood pixel by pixel of the image to be segmented, and the mean value is:

[0033]

[0034] The standard deviation is:

[0035]

[0036] Among them, v ∈ {r, g, b} represents a band of the texture image, ij represents the current pixel position, and w is the neighborhood window diameter;

[0037] (1c) For each pixel position (i, j) of the image,...

Images