Image Comparison
in [Matlab]
|
Prev: Code folding gone?!
Next: dyadic wavelet toolbox
From: Rajesh Sarkar on 23 Apr 2010 10:20 Dear Sir Thank you very much for your valuable ideas. Now I am understanding that I am going into wrong direction and putting worthless questions. But Sir, as I want to compare two images - I want your advice that from where I should start - like what are the aspects of a Black & White image that can be compared? What are the different inbuilt functions of MATLAB (just give some examples) that I can take? Sir, I am very new in MATLAB and I'll be very much grateful if you please help me.
From: Walter Roberson on 23 Apr 2010 12:29 Rajesh Sarkar wrote: > Dear Sir Thank you very much for your valuable ideas. Now I am > understanding that I am going into wrong direction > and putting worthless questions. But Sir, as I want to compare two > images - I want your advice that from where > I should start - like what are the aspects of a Black & > White image that can be compared? What are the different > inbuilt functions of MATLAB (just give some examples) that I can take? > Sir, I am very new in MATLAB and I'll be very > much grateful if you please help me. I see you are immune to patient instruction, even going so far as to email me directly with a word-for-word repetition of your question after I had taken the time to give step-by-step instructions for you to find the answer for yourself. Very well: here are some *specific* image comparison functions: accumarray acos acosd acosh acot acotd acoth acsc acscd acsch addtodate all amd and angle any arrayfun asec asecd asech asin asind asinh atan atand atanh aufinfo aviinfo For a more extensive list, see http://www.mathworks.com/access/helpdesk/help/techdoc/ref/funcalpha.html Don't bother coming back again and asking what "aspects" of the images that these functions compare, as it is _your_ homework and we require that you do _some_ of the thinking.
From: ImageAnalyst on 23 Apr 2010 17:17 Why don't you just post your images and tell us what you're after? No one compares images for the hell of it. You must want to do it for some reason, like you're trying to detect scene changes (e.g. a new truck or building on the site), or you're trying to evaluate compression/decompression algorithms, or whatever. There is no way for us to tell unless you show us and tell us.
From: Walter Roberson on 23 Apr 2010 17:43 ImageAnalyst wrote: > Why don't you just post your images and tell us what you're after? No > one compares images for the hell of it. They do if it is a homework assignment. Especially one that doesn't even ask them to compare any images, just to describe how images can be compared. > You must want to do it for > some reason, like you're trying to detect scene changes (e.g. a new > truck or building on the site), or you're trying to evaluate > compression/decompression algorithms, or whatever. Yup, that old irrelevant thing that I referred to as "the intent of the comparison". My guess is that the original poster would be perfectly happy if he were fobbed off with a couple of functions that are able to do process 2D arrays into a single number (preferably for him), or into a vector of numbers that the same function could be applied to in order to reduce into a single number. And then the original poster would compare the two numbers that came out, and _somehow_ that would be considered to give some kind of percentage "match" between the two images. For example, I bet he'd be happy with this function: 100 * ((sum(P1) / P2) / (sum(P2) / P1)) or 100 * ((P2 \ sum(P1,2)) \ (P1 \ sum(P2,2))) What does it compare? Dunno. But it produces a single percentage as output, and it *looks* like it ought to have some deep meaning for some purpose or other.
From: Rajesh Sarkar on 23 Apr 2010 23:30
Walter Roberson <roberson(a)hushmail.com> wrote in message <hqshtl$l3f$1(a)canopus.cc.umanitoba.ca>... > Rajesh Sarkar wrote: > > Dear Sir Thank you very much for your valuable ideas. Now I am > > understanding that I am going into wrong direction > > and putting worthless questions. But Sir, as I want to compare two > > images - I want your advice that from where > > I should start - like what are the aspects of a Black & > > White image that can be compared? What are the different > > inbuilt functions of MATLAB (just give some examples) that I can take? > > Sir, I am very new in MATLAB and I'll be very > > much grateful if you please help me. > > I see you are immune to patient instruction, even going so far as to > email me directly with a word-for-word repetition of your question after > I had taken the time to give step-by-step instructions for you to find > the answer for yourself. > > Very well: here are some *specific* image comparison functions: > > accumarray > acos > acosd > acosh > acot > acotd > acoth > acsc > acscd > acsch > addtodate > all > amd > and > angle > any > arrayfun > asec > asecd > asech > asin > asind > asinh > atan > atand > atanh > aufinfo > aviinfo > > For a more extensive list, see > http://www.mathworks.com/access/helpdesk/help/techdoc/ref/funcalpha.html > > > Don't bother coming back again and asking what "aspects" of the images > that these functions compare, as it is _your_ homework and we require > that you do _some_ of the thinking. I think I'm irretating you. I want to compare an original signature image with one fraud signature image. Both the signatures were scanned through a scanner. Now the output of my program should be like that the fraud signature is __?__% near to the original one. |