electronic nose circuits

It's not quite that simple-minded. Each dot represents a "dimension" of the metabolite mix produced by the unknown, sorta like coordinates in a multid imensional space, duh. There is an added complication of time there, duh. T he goal is the acquisition of a metabolic signature consisting of time-depe ndent molecular concentrations. Preliminarily, the false negative performan ce appears to be way too high for clinical use. Nobody can work with number s in the 25% range.

is

the

he metabolite mix produced by the unknown, sorta like coordinates in a mult idimensional space, duh. There is an added complication of time there, duh. The goal is the acquisition of a metabolic signature consisting of time-de pendent molecular concentrations. Preliminarily, the false negative perform ance appears to be way too high for clinical use. Nobody can work with numb ers in the 25% range.

Thanks Fred, I think our panel of ecxperts is missing the point. The "goo" on the input is certainly important, but medical diagnostics magic has happened as a result of the readout, primarily. (going out on a limb here). So if you take analysis of breath gases to diagnose cancer, there are various standard modalities o f goo. FTIR is very popular, but you have the eNose, IMS, and others. To summ arize, it is neither the measurement nor the readout which is stunningly im portant here. Instead, there is something else going on here which is right in front of o ur noses. I'll allow Phil to say what that is.

JB

snip

CR's yeah yeah tbird

Thanks Fred, I think our panel of ecxperts is missing the point. The "goo" on the input is certainly important, but medical diagnostics magic has happened as a result of the readout, primarily. (going out on a limb here). So if you take analysis of breath gases to diagnose cancer, there are various standard modalities of goo. FTIR is very popular, but you have the eNose, IMS, and others. To summarize, it is neither the measurement nor the readout which is stunningly important here. Instead, there is something else going on here which is right in front of our noses. I'll allow Phil to say what that is. It may be a major scientific breakthrough. Hint: It's not the measurement. jb

JB

lot

describing

Chinese

For

it's

evil,

quite a

unique

sleep-walking,

multiple

they?

of the

solution.

is.

speckle

differentiation

there,

philosophy

of the

pattern

averaging

you don't

them?

In general my approach to something like would be to minimize the pre-nn signal processing other than bandwidth controls. Chemosensors tend to be rather slow (seconds of response time). Analysing an array of slightly dissimilar sensors requires a training template derived from existing empirical data. This is a neural network property. In all use cases i would have a tendency to go digital as early as possible with each signal source. This makes everything more manageable with already standard algorithms. YMMV

?-)

pattern

if

for

everwhere.

synapses

Well noted. Of course ancient Bharati (Indian) and Chinese sources (and probably Japanese) show this even longer ago.

?-)

Maybe, but not "of course". Before the Pythagoreans, ~700 BC? Links?

Cheers

Phil Hobbs

Of course it's important, it just isn't hard. People have done surface acoustic wave, optics, conductivity, you name it. It's easy.

Cheers

Phil Hobbs

He was more accurate than that - he was out by around 16%. It turned out that incorrect assumptions about the distance and direction between the two towns he used (the distance was measured by camel), combined with inaccuracies in his measurement, happened to give an accurate result by luck.

The mathematics behind his calculations were correct, so it was a great achievement even if the result was by luck.

good

(and

Been too many years since i looked at that last. Before Guahatma Budda tho.

?-)