How is it possible to compare 2 images stored in a byte[] array..??

[srinivasreddy501]

Hi,
i am vinay, sir i am capturing image on mobile for every 5 seconds into byte[] array...
So here for every 5 seconds a image is stored into array...

From that byte[] array, i should be able to compare the 1st image data to the 2nd image data, 2nd image to 3rd image and so on....
so how could be this done programmatically.....
If the above is not possible tell me

1. How can i read them separately from byte[] i.e one image after the other..??
2. How many buffers should i take for this..... should i store each image in a
single buffer..[ 1st image in one buffer, 2nd image in another buffer and compare
both]??

I really beg you help sir please find a way for this sir....

[grahamhughes]

Well, if you want to compare two byte arrays, you need to have two byte arrays to compare. So, you need at least two buffers.

The byte arrays you have are probably JPEG data streams. I'm not sure how useful it is to compare them in this format. You might be better off to have the raw pixel data.

int width = img.getWidth();
int height = img.getHeight();
int[] pixels = new int[width * height];
img.getRGB(pixels, 0, width, 0, 0, width, height);

Each int in the array is a 32 bit ARGB value for a single pixel.

You can compare to int[] easily:

private static boolean areIdentical(int[] a, int[] b) {
    boolean same;
    if (a.length != b.length) {
        same = false;
    } else {
        same = true;
        int i = 0;
        while (same && i < a.length) {
            if (a[i] != b[i]) {
                same = false;
            }
            i++;
        }
    }
    return same;
}

However, this will tell you only whether or not two images are identical. A slight fluctuation in lighting would cause even the smallest difference, and the two images would not be identical (even if they look the same to human eyes).

You might want to think about how much difference between two images you want to detect.

Graham.

[srinivasreddy501]

Well, if you want to compare two byte arrays, you need to have two byte arrays to compare. So, you need at least two buffers.

The byte arrays you have are probably JPEG data streams. I'm not sure how useful it is to compare them in this format. You might be better off to have the raw pixel data.

int width = img.getWidth();
int height = img.getHeight();
int[] pixels = new int[width * height];
img.getRGB(pixels, 0, width, 0, 0, width, height);

Each int in the array is a 32 bit ARGB value for a single pixel.

You can compare to int[] easily:

private static boolean areIdentical(int[] a, int[] b) {
    boolean same;
    if (a.length != b.length) {
        same = false;
    } else {
        same = true;
        int i = 0;
        while (same && i < a.length) {
            if (a[i] != b[i]) {
                same = false;
            }
            i++;
        }
    }
    return same;
}

However, this will tell you only whether or not two images are identical. A slight fluctuation in lighting would cause even the smallest difference, and the two images would not be identical (even if they look the same to human eyes).

You might want to think about how much difference between two images you want to detect.

Graham.

Hi sir,
thanks for ur valuable reply....
Sir how can we know that an image is in which format...
U are saying that i am taking a jpeg image for comparision which would not work for me...
So please tell me how can i get raw pixel data....Sir mention it programmatically sir please....

>>**Sir my need is "I should get raw pixel data after capturing an image from mobile camera".....Sir i am new to this image processing technology so please dont disappoint me sir...give me any code which worths sir....

[grahamhughes]

tell me how can i get raw pixel data

I already did...

You might be better off to have the raw pixel data.

int width = img.getWidth();
int height = img.getHeight();
int[] pixels = new int[width * height];
img.getRGB(pixels, 0, width, 0, 0, width, height);

Each int in the array is a 32 bit ARGB value for a single pixel.