algorithm - How to segment with false edge in in image? -
I try to solve the problem of false edge in image segmentation. See its definition in the figure. The problem of wrong edge is often in the unique image, especially for the medical image. I am getting the method or facility to solve it and to gain a real edge. I try to apply some current method examples: The graph is cut but they are not effective. I also apply shield or entropy here. Can you suggest some idea or feature or method to solve it? You can download the test image here. Note: The original image is "input.mat". Its type is 16 bit
Let us, for example, in my marked image, B is the real edge. We analyze only one point at one point, more intensity than the object area, but it is not so big to assume that the intensity of a point is 300, in which the object The intensity of is approximately 260-280 (humanity property). So we can say that there is a point background, if we calssify it in two areas. Sorry, you are right, the question is that this is a true edge and a false lead (because the other is created by the other region and the background); Created by the right edge object area and background
This issue comes under threshold question. The "false" edge is an edge (visually impaired) Do you have an option to filter the incoming image? A Binarize filter (by changing all the dark brown points black and all the light brown points) will result in a very clear edge. Applying the filter to find the edge of the technique, show the graphics on the original image.
Another technique is to change your step threshold. Assuming that the difference between dark brown and light gray is less than the difference between dark brown and background.
Can you change the input image? Creating a backlit image will not be able to detect surface flaws.
Hope this gives you some ideas.
Comments
Post a Comment