Then again, there's all kinds of psycho-visual effects that can screw up this kind of research without even taking this into account. In other words, if your data isn't lining up, the fundamental problem is that the response of individuals' eyes are rarely comparable. Not only are they highly non-linear, the only reason they appear to be smooth is that they are a statistical average of many, many different individuals' efficiency curves. These are the conversion factors that take into account that the eye has widely varying response to different frequencies of light. Converting between the two involves the use of what are called Spectral Luminous Efficiency Curves. In reality, we see on a different, completely NONLINEAR scale, known as the photometric scale (Talbots, lumens, etc. The biggest problem is that any mathematical computations about modelling light are done in radiometric space, where light is considered pure energy of varying frequency (Joules, Watts, and other simple(?) units). This is probably exactly your problem, but even so, I feel your research is far to subjective without adding quite a bit of statistical research (which has already been done, so don't completely reinvent the wheel). Color perception is decidely non-linear and doesn't even fit a nice equation. You are seeing the relative sensitivities of your eye to pick up colors. This is the reason you see odd patterns in color changes in a rainbow. You are going to get 'dead areas' of the model where the eye is not so good at viewing differences, and active areas where in a small space there seem to be quite a few different colors because the eye is perceptually sensitive in that area. So if you are trying to do a 3-space transform and distance and finding that colors which are linearly close to each other in CIELAB space are perceptually very different, it is very likely because it is a color in a space where the human eye is more perceptually sensitive. Thus if you are trying to determine at what point the eye will say 'hey there is a block of a different color on top of that one', blue is going to be an important part of your color model since while your eye isn't so good at picking up how bright a blue is, it is very sensitive to the relative shade around the blue wavelength something is. The eye is most sensitive to chrominance changes in BLUE. The eye is most sensitive to luminance changes in GREEN. This is a complex problem and I think a lot of the answers I'm reading on this topic center around a misconception on a lot of people's part about the difference in luminance perception and chrominance perception.

Cielab software work