0
$\begingroup$

It is common to say that the YUV color model is inspired from the human visual system (HVS), knowing that we have many more rods than cones, and consequently, we are more sensitive to light than color. However, it does not seem to be that simple. Rods become active under low light conditions, and cones are not totally insensitive to light. These details seem to be ignored by the literature. My question is: is it reasonable to build HVS models for image processing only based on the number of rods and cones?

For instance, in the paper "A new color image quality measure based on YUV transformation and PSNR for human vision system" by Yalman and Ertürk (http://journals.tubitak.gov.tr/elektrik/issues/elk-13-21-2/elk-21-2-20-1111-11.pdf), a new quality metric called CQM is proposed. It consists in computing separated PSNRs on the Y, U and V channels, and combining them with a weighted average, whose weights are adjusted according to the number of rods and cones. But, once again, is it reasonable knowing that rods become active under low light conditions?

$\endgroup$

bumped to the homepage by Community 2 days ago

This question has answers that may be good or bad; the system has marked it active so that they can be reviewed.

0
$\begingroup$

As far as i know the YUV color space was not inspired from HVS, but from technical compatibility reasons; to transmit new color video television broadcast signals over the existing black and white infrastructure.

Essentially, the Y component of a color image (which is captured naturally in RGB system) carries the black and white or the brightness information of the image and is transmitted in such a way that old TV equipments (those that do not have color receiving capabilities) can still receive and decode the Y component as if it were the old black and white signal and still be able to display it without any hardware modification.

While at the same time, new (back then in 50s, 60s) color TVs were capable of also receiving transmitted U and V components and were able to display color pictures by properly decoding the YUV signal using some sort of comb filtering in NTSC and PAL systems.

Furthermore, Y carries more bandwidth and information than U and V. There's a separate color carrier towards the end of the channel bandwidth (before the audio carrier).

Because of certain industrial standardization concerns, YUV and its variants (YPbPr, YCbCr) achieved wide range of acceptance and being used in almost every image and video electronic equipments and codecs (in its digital form).

It's true that Human sensititivity to color and brightness are different due to cone and rod structures. And researchers can take advantage of this information to optimize their metrics with respect to human visual system. The paper most probably assumes such a correlation.

$\endgroup$
  • $\begingroup$ Thank you for the clear explanation on YUV! However, regarding the metric, is it relevant to take advantage of rods/cones distribution when working on printed images or backlit monitors? Especially when we know that rods become active in low-light conditions. $\endgroup$ – benlaug Jan 15 at 13:30

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.