Semi-automatic image cutting method based on nuclear transfer
An image segmentation and kernel transfer technology, applied in the field of image processing, can solve problems such as inability to maintain the consistency of superpixel data points, unclear edges of objects of interest, and unlearned image data, etc., to achieve consistency and false match rates Low, the effect of improving accuracy
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Publication Date
- 2012-09-12
- Estimated Expiration
- Not applicable · inactive patent
Smart Images
Figure 1 Figure 2 Figure 3
Abstract
Description
technical field
[0001] The invention belongs to the field of image processing, in particular to a semi-automatic image segmentation method based on kernel transfer, which can be used for target extraction and target detection. Background technique
[0002] Image segmentation is one of the most basic problems in computer vision and digital image processing, and it is the basis for further analysis, recognition, tracking and understanding of images. Image segmentation divides the pixels in the image into different subsets according to different visual features or semantics. Specifically, image segmentation is the process of assigning a label to each pixel in the image, the purpose of which is to make pixels with similar visual features or the same semantics have the same label. At the same time, it is of great significance to study image segmentation, which can be widely used in many fields such as semi-automatic image retrieval, video conferencing, 2D to 3D technology, targe...
Examples
Embodiment Construction
[0030] The specific realization and effect of the present invention are described in further detail below with reference to the accompanying drawings:
[0031] refer to figure 1 , the implementation steps of the present invention are as follows:
[0032] Step 1. Input an image and pre-segment it into a set S of superpixels.
[0033] First, input an image, and use the mean shift algorithm to pre-segment the image to obtain the label of each pixel;
[0034] Then, a set of pixels with the same label is called a superpixel, and a set of n superpixels is obtained, which are respectively identified as s i , i=1, 2, ..., n, to obtain the superpixel set of the pre-segmented image
[0035] where s i is the i-th region obtained by pre-segmenting the input image by the mean shift method, and n represents the number of superpixels contained in S.
[0036] The mean shift method can be found in literature: D. Comaniciu, P. Meer, "Mean shift: a robust approach toward feature space analy...