From what I see/measure....the main reason the incans look brighter sometimes is due to the generally tighter beam profile. Essentially, its the same reason people will look at the beam from a 900 lumen SC600 against a wall next to a 131 lumen mag, and state the 131 L mag is brighter...the tighter beam causes more glare, which the eye tells the brain is "bright".
As incans can typically emit from a smaller point source, they do tend to have throw more easily in many scenarios. The reflectors can be shallow but wide, and throw well, whereas an LED light requires a proportionally deeper reflector to get throw, while still needing diameter to achieve their best throw.
This can mean that the form factors can lend themselves to incan's, HID, etc, having intrinsic throw advantages.
When sighting targets, the difference in light quality, as far as target acquisition, motion tracking and aim accuracy, is not actually that relevant. The lux on target seems to be far more relevant than the tint or spectrum itself...as long as the spectrum includes yellow/green, where the eye is optimized. Red light was the hardest for fast targeting, especially for people with low red sensitivity. Lux trumps color though, very consistently.
When incans look dimmer, its where they have a floody beam, reversing the effect, as now their beam is more spread out and diluted, etc.
As for pattern recognition, the incans are better for spotting things like a dark clothed person in shadows, as the subtleties of close color patterns can stand out better in most cases. The LED can do very well in some cases, but, overall, the incans do have an edge. When the target is wearing camouflage, ironically, the LED can do better under some conditions, as the patterns meant to disrupt the outline of the target can be accentuated, defeating the process. This is akin to when we use different wavelengths in general, as a pattern/textile combination designed to concealing when viewed with night vision, or day light, may be revealing under UV, etc.
So, lumens are lumens, and if an IS measures 2 lights as 100 lumens, they both emit 100 lumens. If one 100 L light has a cd of 10k, and the other has a cd of 50k, most people will say the 50k cd 100 L is brighter than the 10k 100L.
Lumen measurements INCLUDE a measurement across the visible spectrum, so, if a light's output is MISSING part of the spectrum, its lumen out put is proportionally lower. As an LED's frequency emission is not as evenly distributed as an incan's, the LED will be proportionally lower rated when little or no signal is received in some wavelengths, etc. The LED may emit most of its light in certain wavelengths, with stronger peaks in those wavelengths. If the peaks are tall and broad enough, and include the wavelengths that the lumen measurement is weighting heavily, (The lumen measurements weight the measurement towards what the eye can perceive best), it can result in a beam that is strong where the eye can use it best.
So, overall, in my use, etc, the main advantage to the LED lights are the lower power consumption/smaller form factors relative to output that can be achieved, so a smaller light with a longer run time can accomplish the same task normally reserved for an incan. The quality of light for most functions works fine, and I simply use incan/HID/LED lights where the advantages/disadvantages of each make sense.