Way of Calculating NPCR and UACI

Can someone clear the doubt about way to calculate NPCR and UACI as some suggests to change one or more bits of each pixel value in image while otther prefers to change single pixel value in whole image and also some do by a change of single bit in key, Which is the correct way to evaluate these values?

Answer 1

[Saurabh]

Hi Kajal,

I found one IEEE research paper that speaks about NPCR and UACI. The paper states that the NPCR and UACI tests changing pixels and the number of averaged intensities between plaintext images is subtle that is usually a single pixel.

The link to the official documentation page of the IEEE research paper:

https://citeseerx.ist.psu.edu/document?repid=rep1&type=pdf&doi=2b479abce221135af6065f9f8352e09cbfb5733a

If you want information on how to calculate NPCR and UACI between two encrypted versions of the same image, different only by one pixel, You can refer to the below file exchange post, which is written by one of the community member:

https://www.mathworks.com/matlabcentral/fileexchange/155712-npcr-and-uaci-measures-for-image-encryption?s_tid=srchtitle

I hope this was helpful.

[Kajal Chawla]

a single pixel means that pixel value, wherever is repeated , will also be changed, or it means just at one location pixel value is changed?

[Umar]

Hi @Kajal Chawla,

Altering a single pixel involves updating the intensity or color value at a precise coordinate, rather than globally changing all pixels with the same value. Please let me know if you have any further questions.

[Kajal Chawla]

if a single pixel is changed ,then pixel permutation will change the pixel value at just one position further diffusion process also affect only that changed pixel then how NPCR is expected to be close to 99....i mean how it can affect all the other pixels in encrypted image?

[Umar]

Hi Kajal Chawla,

Even if only one pixel is modified, the diffusion process in encryption algorithms ensures that this change propagates throughout the image, affecting the neighboring pixels due to the interdependence of pixel values. Therefore, despite altering a single pixel, the diffusion mechanism causes a ripple effect, leading to a high NPCR value close to 99 as the change influences a significant portion of the image.

[Kajal Chawla]

Thank you....This helped a lot.

[Umar]

Hi @ Kajal Chawla, no problem, glad to help out. Please let us know if you have any further questions. We will be more happy to help you.