Target Detection

Detects a Target image in a given image and focuses on it.

Description

This program detects a target image in a given image according to maximization of Cross Correlation.

Sample Run:

>> target

 This program detects a target in an image
 Entering the image for MATLAB...
 Save the image or its copy in MATLAB working Directory
 Enter the name of the image file (filename.ext) : myphoto.bmp

 Entering the target image for MATLAB...
 Save the target image or its copy in MATLAB working Directory
 Enter the name of the target image file (filename.ext) : eye.bmp

 Do you like to SAVE Result Image? (y/n) : n

 You choose NOT to SAVE the Result Image

• Required Products: Image Processing Toolbox
• MATLAB release: MATLAB 6.0 (R12)

[Mina Saleh]

 

[hennesa hennesa]

 

[C Dose]

Very Well Done!!!

[Shahzad Ahmad]

 

[abhay kulkarni]

well prepare

[farzad mazaherian]

please send M-files of the matlab for image processing
best wishes

[Tashfeen Karamat]

Good work.

[Abdul Sultan]

Nice Work.

[yanwan yan]

Very good

[Changchaiyo Wattana]

I think,it's a good programm. Now,I am testing this programm and I feel good . If you have programm which can detect other part of object , I give a programm to test again please . Thank you very much

[Kwok Denis]

Super cool!!!!

[masoud saadatkhah]

 

[lissa anastasia]

tq for the good programming.i just want to know is that the same method in this program for me to detect the car plate number as my target?

[Bob Schafer]

Awesome, and simple! Should advertise that it can also find a *similar* target: for example, show it a scene with a face in it, give a different face as target, and it'll (often) find the face in the scene.
Can change the line
pixel = find(output == 1);
to something like
pixel = find(output >= thresh);
to find all instances of a target (or places similar to the target, e.g. all faces).

[Siddharth Joshi]

Way to go!.... a very useful application...
Used it for my project...

[kratos kong]

Good....i am currently doing research on teleoperated bomb disposal robot vision system...i wonder if i can use this features or any other features in matlab image processing to enhance its peformance..Can someone give me any idea? thanks. ([e-mail])

[aziz osman]

Ita good, but it tries to give a positive output eventhough the images do not match.That could pose a problem i believe.

[Chon Ji]

I can't make it work, I tried to input 2 bmp files but it shows an error
conversion to cell from double is not possible.

Error in ==> test at 53
mask(:,:,1) = m;

How can I make it work?

[siyao fu]

need to have a pre-definded pic first, use it to travese all the image to match the one.

[Fernando Castro]

Very good!!!

[Harold Odom]

 

[nanda kishore velaga]

i find it interesting....iwant to prepare documentation for this.........please help me.....what topics i can include in it ?.....please help me

[hana qudah]

 

[rajarao chatla]

it's working but im unable to follow this detection part. Pls help

[ABHAY PRATAP]

 

[arnold jansen]

i need source code about face recognition using gabor wavelet.would you help me please....

[mamadoe86 mamadoe]

good work

[Lin Zhang]

This is a template matching topic.

I would like to know how to deal with the rotation and scaling changes for the template in the input image.

When there is no rotation and scaling change, the cross-correlation method can do this detection job. The cross-correlation method can not detect the template when it is rotated and changed.

I am looking forward to anyone's response. Many thanks in advance.

[jingkun guo]

good!

[raffaele pappaterra]

 

[Ashish Pant]

The implementation above is good but seems to fail on real world images extracted from a video sequence. Due to intensity flicker and motion of even the smallest objects, targets dont get detected on subsequent images.Further more the size of target image is crucial since in a real world sequences areas within an object can appear to be similiar and thus lead to multiple area detection as i am obtaining. With additional work this code could overcome these problems.

[majid chababi]

indexed images by color,textur and forme

[kum jinlong]

good

[Ashar Malik]

the effort is good yet the problem of matching is not completely solved, the rotated or tilted objects create a real mess and you have to permute, which takes centuries .... well at least for now it does....
/////
i am working on something related and as soon as i get it up will post it here ....

[Alberto Cantillo]

 

[flame metal]

I am just a beginner in image processin.can someone please tell me what is the function of following codes in this program? I will be really grateful.

1. output = (1/o)*out;
2. pixel = find(output == 1);
3. pcolumn = fix(pixel / sizw(1));
   prow = mod(pixel,sizw(1));
   rdis = fix(sizt(1)/2);
   cdis = fix(sizt(2)/2);
4. m = 0.5 * (m + 1);

[WTF WTF]

 

[Jun wan]

very done,thanks

[Raymond Cheng]

Thanks for your sharing.

[Sisir Kanti]

Hi, i didnot understand this part. Please help!

m = 0.5 * (m + 1);
mask(:,:,1) = m;
mask(:,:,2) = m;
mask(:,:,3) = m;
final = mask .* w;

[Mohamed]

ya ! cool one ...................,

[sahil]

can any1 plz xpalin me the code inside the {} brackets..
{pixel = find(output == 1);
pcolumn = fix(pixel / sizw(1));
prow = mod(pixel,sizw(1));
rdis = fix(sizt(1)/2);
cdis = fix(sizt(2)/2);
cmin = pcolumn - cdis;
cmax = pcolumn + cdis;
rmin = prow - rdis;
rmax = prow + rdis;
c = [cmin cmin cmax cmax];
r = [rmin rmax rmax rmin];
m = roipoly(ww,c,r);
m = im2double(m);
m = 0.5 * (m + 1);
mask(:,:,1) = m;
mask(:,:,2) = m;
mask(:,:,3) = m;
final = mask .* w;
}
i need to understand it as soon as possible..

[sahil]

 

[Chandra Kurniawan]