double derivative of an image

hello,

i am working on an image in which i need to find out the inner most points of it. can i apply the concept of laplacian operator to find the same?

and please help me even if there is any other method with the help of which i can find them.

thank you.

[Jan]

Please start with an exact definition of "innermost points". Does the margin paly a role? How to handle multiple "innermost points" e.g. if the input is a circle or a hexagon? Why are the red marked points the solution and not these green ones:

[Meghna Chaudhary]

innermost points are the points that lie in the concave portion of the image (highlighted with red circles).

red circles in here are the points i want to extract. this is not the result.

is there any specific code i can use to obtain the same and avoid any other point.

Answer 1

[darova]

What about manually drawing line?

clc, clear

I = imread('original image.png');
I = rgb2gray(I);
I = im2bw(I);
imshow(I)

h = msgbox('Draw line');
waitfor(h)
line = imline;
xy = line.getPosition;
L = sqrt( sum(diff(xy).^2,2) );

[Meghna Chaudhary]

thank you darova,

but i can not plot them manually, because there is a series of images in which i have to apply the same algorithm.

[darova]

• use bwboundaries()
• reduce points to get a bit smoother curve
• find angle between neighbour segments

but how to choose "right" points?

see attached script

[Meghna Chaudhary]

thank you darova.

[darova]

you are welcome