I have attached the image here, in this, the black color region is background, white region is region of interest (roi). red points are inside the region and yellow points are outside the regions. Programmatically (using matlab code), how do i find whether a given point is inside or outside the roi. (color of points can be ignored, it is used just for explanation).
Thanks in advance.
Hi,
I'm not an image processing geek, but if it can help you :
You could cluster your regions of interest (if it not already the case) with something like k-means. You could then process a hull (concave or convex, according to what you mean by "inside" the region) for each cluster, and check if the point is inside one of them.
Mickael
https://en.wikibooks.org/wiki/Algorithm_Implementation/Geometry/Convex_hull/Monotone_chain
http://www.mathworks.com/matlabcentral/answers/33194-point-in-convex-hull
http://nl.mathworks.com/help/stats/kmeans.html
http://nl.mathworks.com/help/stats/cluster-analysis.html
Dear Deepa,
Convert RGB images into HSV color space, and do segmentation using LVQ neural network depending on the hue values of each pixel and size of the image.
All the best...!
Ray casting and winding number algorithms determine whether a point lies inside or outside a region. When a point lies exactly on the boundary - or within a tolerance ε - colinearity is checked. For a point on boundary, the colinearity test is positive for edge-defining points.
Hi Deepa,
I cannot recall if you downloaded MIPAR, but here is a recipe which accomplishes what you need. Even if you do not wish to ultimately use MIPAR for your processing, loading this recipe will reveal algorithm logic that may assist you in writing your own code.
- I have also attached a grayscale representation of your image, and the outline segmentation output by the recipe (GREEN = ROI, RED = Points Inside, YELLOW = Points Outside).
- Finally, I have attached a brief screen recording demonstrating the recipe
Let me know of any questions!
Cheers,
John
Hi Deepa,
Do you only want to know whether points are inside the regions of interest or in the background? If so, you can just check the intensity of the pixel where the points are in the binary image and if it is 1 then is inside otherwise is outside.
I would probably use imfill(binary_image) first, to remove the spurious black dots in the ROi. And then, bwconncomp(binary_image) to extract only the regions with large number of white pixels in order to remove the white dots in the background.
Hope it helps.
Hi Deepa,
If u use black&white images, u can use structure element (SE) to find your interest region. If the SE was in the region, u can place center of SE as interest point and if SE was out of region, u can center as interest point.
For example, u select 3*3 kernel as SE that all of element's are one. Now you move SE on your image, if product of SE by region of image that SE on it, is 9 (if interest region is white), it can your interest region.
Hope it helps.
Please check this link :
https://in.mathworks.com/help/matlab/ref/inpolygon.html
hi
Try following:
1. Obtain image matrix say A
2. Convert A to B where B is matrix obtained by thresh holding
A into regions of interest , so B will have blocks of submatrices
with different numbers(maybe region numbers)
3. Let C = matrix of points to be determined set to one and other elements set to zero.
4. Obtain D = C. *B (note elemental multiplication) . this matrix
is of interest to you.
Cheers
I have an Idea. If you know the coordinates of closed boundary , you can initialize a binary matrix with a white back ground and black boundary. After that, you can apply flood fill algorithm and fill the closed boundary using 'imfill' command in MTLAB. After that, you create a 2D loop and compare whether each pixel coordinate has a black or white pixel. If it is black, then the coordinate is inside the region otherwise it is in outside region.
Let A is your binary image where the boundary only has black pixels and all boundary pixels are 8- connected to the subsequent pixels.
A1=imfill(A,'holes');
for i=1:size(A1,1)
for j=1:size(A1,2)
if A1(i,j)==1
disp('It is an outside region')
else
disp('It is an inside region')
end
end
end
%% If you have any issue you may reply
- Badges
- Science topic