i am currently doing(going to start) a new project on image comparison method. Basically i want to compare 2 images changed under any circumstances( moustache,beard,inclination etc). I will be going to compare the new image with the database i have and give the closest possible match i get. I request you to help me with this. pls forward me whatever material or research papers you have in relation to this project.I will be very thankful to you.

 Réponse acceptée

Image Analyst
Image Analyst le 25 Fév 2012

0 votes

You could use SSIM : http://en.wikipedia.org/wiki/SSIM "The structural similarity (SSIM) index is a method for measuring the similarity between two images. The SSIM index is a full reference metric, in other words, the measuring of image quality based on an initial uncompressed or distortion-free image as reference. SSIM is designed to improve on traditional methods like peak signal-to-noise ratio (PSNR) and mean squared error (MSE), which have proved to be inconsistent with human eye perception."
or better yet, do some research on CBIR. Just Google "CBIR database"

14 commentaires

sahil316
sahil316 le 25 Fév 2012
how i would deal with inclination etc???does this method include that??
Image Analyst
Image Analyst le 25 Fév 2012
I have no idea what inclination means in the context of CBIR or image comparison.
sahil316
sahil316 le 25 Fév 2012
like if the image has been titled or rotated by some angle
Image Analyst
Image Analyst le 25 Fév 2012
Then you're best off using something like SURF (see wikipedia) which has been used rather successfully to find matching images in databases that have been cropped, scaled, and/or rotated.
sahil316
sahil316 le 26 Fév 2012
thanx a lot..i am going to study that.
sahil316
sahil316 le 11 Mar 2012
sir do you know anything about face detection.Like if i have to extract a face out of an image and then i can compare it with the database i have using SURF.
Image Analyst
Image Analyst le 11 Mar 2012
Why don't you use face detection built in to the 2012a Computer Vision Toolbox: http://www.mathworks.com/help/toolbox/vision/gs/btd0kx6-1.html
sahil316
sahil316 le 12 Mar 2012
Sir i dont have this toolbox..is there any other way to do that using the other toolboxes??
sahil316
sahil316 le 14 Mar 2012
sir is there any other way in matlab for the same???
Image Analyst
Image Analyst le 14 Mar 2012
Yes, if you write it yourself, or find something in the File Exchange. But I don't have any face recognition/detection code for you.
sahil316
sahil316 le 15 Mar 2012
all right sir.i will look into it.thanx again!
sahil316
sahil316 le 20 Avr 2012
sir can you please give me the list of all the face detection methods which can be applied using matlab.Actually now i am going to compare all the methods in my project and give conclusion on which one is most optimal.
Image Analyst
Image Analyst le 20 Avr 2012
Here is the list: http://iris.usc.edu/Vision-Notes/bibliography/contentspeople.html#Face%20Recognition,%20Detection,%20Tracking,%20Gesture%20Recognition,%20Fingerprints,%20Biometrics
sahil316
sahil316 le 21 Avr 2012
sir i did get this list from your comment on some other link but i am unable to find source codes in matlab in this website. DO u have any links of matlab codes with you that can be of my help.??thanx in advance!

Connectez-vous pour commenter.

Plus de réponses (1)

Ali
Ali le 25 Fév 2012

0 votes

You could try computing either the mean square error or signal to noise ratio of the two images and see which one gives you the least errors.

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by