Top point detection in an image

[minuri]

I am doing a project on multiscaling and I need to detect the top points. For each Gaussian filtered image the points are to be calculated. Can someone suggest methods for the identifiation of top points in an image?

 

[lordsathish]

what you mean by top points of the image does it mean the maxima of the image... if you want to detect the local maxima then try using the laplacian filter on the image... i.e convolving the image with the laplacian mask... as you have said that you are using a gaussian filter before this it is better to use the laplacian of gaussian filter on the image... this will reduce the number of computations altogether... in matlab there is a function 'fspecial' that will return the laplacian of gaussian mask and then use imfilter to filter it...

 

[minuri]

It is actually detection of the singular points in multiscale space for the face recognition. It is required that I detect these top ponts and then as the sale is varying the points takes a critical path and by studying that it should reveal whether the point is significant or not as the significant points are said to be preserved inspite of the scale. Do you have any material regarding this implementation? I will be glad if you could send me or tell me where to download.

 

[Langat]

Hi, Am also looking for image reconstruction algorithm using top points and critical points? Especially in R software. Could anyone know please assist