MathLab working on images

I have 2 images: VISUALSALIENCY and EVENTSALIENCY ( both images are in grey scale).. in VISULASALIENCY white is minimum and black is maximum.. in EVENTSALIENCY white is minumum and black is maximum .. i have to do an inversion , and should be somthing using the matrix of ONE ... ONE ...= ONE - VISUAL .. something similiar ....

then i have to do the thresholding of both images.. i need help thanks

Answer 1

[Image Analyst]

No inversion is necessary. Just reverse the sign of the > sign if you want to find maximums that have values the opposite of what you'd normally think they should have.

% Threshold to get dark values, which are max saliencies for both arrays.
maxVisual = VISUALSALIENCY < 128; % or whatever
maxEvent = EVENTSALIENCY < 128; % or whatever
bothMax = maxVisual & maxEvent; % If you're interested in where both are max.

[francesco]

sorry i forget that after i need to do an INTERSECTION between the 2 images , is the reason i need the inversion.. because if i use the intersection and in EVENT i have max and min and in VISUAL i have another max and min i will have a bad result.. can u please give me a code to do the intersection??

my teacher used something with 1-Visual= ??? i dont remember what

[Image Analyst]

That's what I gave you. If you want, you can do

maxEvent = ones(size(VISUALSALIENCY)) - (VISUALSALIENCY> 128); % or whatever

but I don't really know why you'd want to do that instead - it takes twice as long to give you the same result. Why don't you just post your images and say what region(s) you want to extract/find.

[francesco]

Thx a lot ( and sorry for my bad english) , so , i have EVENT and VISUAL.. i have to do the inversion of Visual to change min with max in the grey scael , then do the threshold of both images and at end do the intersection between the 2 images.. ( for the intersection i see the images as 2 matrix and i think i can use the .* so where i have 0 0 ->0 0 1->0 1 1->1

[francesco]

[francesco]

[Image Analyst]

What regions in the image do you want to find the intersection of? Wouldn't you just AND the two thresholded images? That's what your somewhat confusing 0 0 1 sentence seems to say, if I interpreted it correctly. Again, no actual inversion is necessary, you simply need to have the right sign, < or >, when you do thresholding.

[francesco]

i have to do the intersecton between the 2 fulls images.. i am not really good on mathlab , ..my teacher said i have to do first the inversion of visualsaliency because max and min are the opposite of event , then do the treeshold of both images and at end do the intersection between the 2 fulls images treshoolded

[Image Analyst]

Can you outline in red the two regions that should be ANDed? And show the final image. Is the final image the house image but only in the pair of really dark rectangles and everything else is black? And I know what your teacher is saying, but again, you don't actually need to do an inversion since it can be handles by how you do the thresholding.

Answer 2

[Walter Roberson]

if the values in an array are in the range 0 to 1, then 1 minus the array will give the greyscale inversion.