What are the traditional image segmentation methods

1. Threshold-based segmentation methods

Gray-scale thresholding segmentation method is one of the most commonly used parallel region techniques, and it is the category with the largest number of applications in image segmentation. Thresholding segmentation method is actually the change of input image f to output image g

Where T is the threshold value; for the image element of the object, g(i,j)=1, for the image element of the background, g(i,j)=0.

It can be seen that the key to the thresholding segmentation algorithm is to determine the threshold value, and if you can determine a suitable threshold value can be accurately segmented the image. After the threshold value is determined, the comparison of the threshold with the gray value of the pixel point and the pixel segmentation can be carried out in parallel for each pixel, and the result of the segmentation directly gives the image region.

The advantages of threshold segmentation are computational simplicity, high computational efficiency, and speed. It is widely used in applications where computational efficiency is important (e.g., for software implementations).

2. Region-based Segmentation Methods

Region growing and split-merge methods are two typical serial region techniques, where the processing of the subsequent steps of the segmentation process is determined by judgment based on the results of the previous steps.

(1) Region growing

The basic idea of region growing is to assemble pixels with similar properties to form a region. Specifically, first find a seed pixel for each region to be segmented as the starting point for growth, and then merge the pixels in the neighborhood around the seed pixel that have the same or similar properties as the seed pixel (determined according to some pre-determined growth or similarity criterion) into the region where the seed pixel is located. The above process is continued by treating these new pixels as new seed pixels until no more pixels satisfying the conditions can be included. In this way a region is grown.

(2) Region splitting and merging

Region growing starts from a certain pixel or certain pixel points and ends up with the whole region, which leads to target extraction. Split-merge is almost the inverse process of region growing: start from the whole image, keep splitting to get each sub-region, and then merge the foreground region to realize the target extraction. The assumption of split-merge is that for an image, the foreground region is composed of a number of interconnected pixels, so if an image is split down to the pixel level, then it can be determined whether the pixel is a foreground pixel or not. When all the pixels or sub-regions have completed the judgment, the foreground target can be obtained by merging the foreground regions or pixels.

3. Edge-based segmentation

Edge-based segmentation refers to the determination of the end of a region, i.e., the beginning of another region, through edge detection, i.e., detecting the gray level or the structure of the place where there is a mutation. Different images have different gray levels, and there are generally distinct edges at the boundaries, which can be used to segment the image using this feature.

4. Segmentation methods based on specific theories

Image segmentation so far there is no generalized theory of its own. With the introduction of new theories and methods in various disciplines, image segmentation methods combined with some specific theories and methods have emerged, mainly: image segmentation methods based on cluster analysis, segmentation methods based on fuzzy set theory, and so on.

5. Genetic coding based segmentation method

Genetic coding based segmentation method refers to the method of separating the image background and target pixels by dividing them regionally, which is represented by different genetic codes. The method has the advantage of fast processing speed, but the algorithm is difficult to implement.

6. Segmentation methods based on wavelet transform

Wavelet transform is a mathematical tool that has been widely used in recent years, which has good localization properties in both time and frequency domains, and the wavelet transform has multiscale properties, which can analyze the signal on different scales, and thus has been applied in many aspects of image processing and analysis.

The basic idea of wavelet transform-based threshold image segmentation is to first decompose the histogram of the image into different levels of wavelet coefficients by the binary wavelet transform, and then select the threshold threshold based on the given segmentation criterion and wavelet coefficients, and then finally use the threshold to mark the region of the image segmentation. The whole segmentation process is controlled by scale changes from coarse to fine, i.e., the starting segmentation is achieved by the histogram projected on the coarse L2(R) subspace, and if the segmentation is not ideal, the image segmentation is gradually refined using wavelet coefficients of the histogram on the fine subspace. The computation of the segmentation algorithm will vary linearly with the size of the image.

7. Neural Network Based Segmentation Methods

In recent years, artificial neural network recognition technology has attracted widespread attention and is applied to image segmentation. The basic idea of neural network based segmentation method is to train a multilayer perceptual machine to get a linear decision function, and then use the decision function to classify the pixels to achieve the purpose of segmentation. This method requires a large amount of training data. The neural network exists huge amount of connections and easily introduces spatial information, which can better solve the problem of noise and inhomogeneity in the image. The choice of network structure is the main problem to be solved by this method.