A Natural Image Superpixel Segmentation Method Based on Graph Model
A superpixel segmentation and natural image technology, applied in the field of superpixel segmentation of natural images, can solve the problems of unbalanced superpixels and high superpixel segmentation accuracy, achieve optimal segmentation performance and improve accuracy
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
specific Embodiment approach 1
[0021] Specific implementation mode one: combine figure 1 Describe this embodiment, a kind of natural image superpixel segmentation method based on graph (Graph) model of this embodiment is specifically called:
[0022] Step 1: Map the input natural image into a weighted graph;
[0023] Step 2: Input the number of K superpixels expected to be generated, and perform uniform grid sampling on the weighted map in step 1 according to the number of K superpixels to obtain the initial positions of K superpixels, and K is a positive integer;
[0024] Step 3: clustering is performed on the basis of the initial positions of the K superpixels obtained in step 2 to generate superpixels;
[0025] Step 4: Optimize the boundaries of the superpixels generated in step 3 to obtain the result of superpixel segmentation.
specific Embodiment approach 2
[0026] Embodiment 2: The difference between this embodiment and Embodiment 1 is that in the step 1, the input natural image is mapped to a weighted graph; the specific process is:
[0027] Gaussian filtering is performed on the original natural image I, and each pixel i in the original natural image I after Gaussian filtering is mapped to a vertex v of the graph G i , obtain the vertex set V; i=1,2,...N, N is the total number of pixels of the original image I, and the value is a positive integer;
[0028] The graph G is a graph Graph;
[0029] For each vertex v in graph G i are in its 8 neighborhoods (such as Figure 5 ) vertex v j Use edge e(i, j) to connect to get edge set E of graph G, j=1, 2, ... N; assign weight to each edge e(i, j) in E, after weighting, e(i ,j) is recorded as w(i,j), and w(i,j) is v i with v j Euclidean distance d on (r,g,b) space spectral Euclidean distance d from (x,y) space spatial The weighted sum of:
[0030] w(i,j)=λ·d spectral +(1-λ)·d ...
specific Embodiment approach 3
[0034] Specific embodiment three: the difference between this embodiment and specific embodiment one or two is: in the step two, input the number of K superpixels expected to be generated, and perform a uniform network on the weighted map in step one according to the number of K superpixels grid sampling,
[0035] The initial positions of K superpixels are obtained, and the value of K is a positive integer; the specific process is:
[0036] Input the number of K superpixels expected to be generated (manually set based on experience), and sample the weighted map in step 1 with a uniform grid separated by s pixels to obtain the starting positions of K superpixels; in order to generate approximately uniform size superpixels, grid spacing K superpixels are denoted as (C 0 ,C 1 ,C 2 ,...,C K-1 );
[0037] Among them, C l is the set of all pixels in the lth superpixel, 0≤l≤k-1.
[0038] Other steps and parameters are the same as those in Embodiment 1 or Embodiment 2.
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More