Using AI for manual color perception tests (Iodine/Iodide test)

[Ameer214]

Hi Everyone,

From a recent ICP test I learned that I needed more Iodine. Long story short, I decided to give the Seachem test a shot. The test is tricky because you have a brief window to detect the color before it clears. I decided to take an image when the color peaked and used AI to analyze it for me. I included a reference in my test as well. The reference was the same as the one that came with the kit. It is suppose to be 0.06mg/L. My ICP test reports a Iodine level of 0.03mg/L. I guess this includes Iodide and Iodine trapped in organics? Anyways, I got promising results with ChatGPT. It correctly reported the reference value and reported my Iodide as 0.01 to 0.02mg/L (see screenshot below).

 

[Lasse]

Clever idea

Sincerely Lasse

 

[Cow]

I’ve noticed in reviews that the Red Sea iodine kit often seems to test lower than people’s ICP results. I’ve been debating whether it’s worth trying a hobby kit at all. I haven’t looked into the Salifert kit yet, but based on cost and accuracy concerns, it almost seems more practical to rely on ICP testing if these kits can’t be trusted.

I found an interesting review and was curious if anyone has any thoughts on dosing differently based off ICP vs Kit tests
"In response to the 1 star review stating,
test kits from Red Sea, Salifert, Elos, Hanna, Nyos, SeaChem, Giesemann, Lamotte, and Hach, and then compared them to results from simultaneous water samples sent to Triton Labs for spectroscopy. Regarding Iodine testing, ALL of the available test kits will give you inaccurate results even when performed correctly. In many cases, the test will read as 0 or low iodine when in fact your level may be dangerously high already.

Triton ICP-OES testing measure atomic concentrations and not molecular concentration. The two types of testing measure different things. The Element "I" takes on many forms in water including including iodide (I–), iodate (IO3–), iodine (I2), hypoiodate (IO–), hypoiodous acid (HOI), and many different organic forms, such as methyliodide (CH3I). ICP-OES breaks these molecules apart and counts the I atoms. Test kits measure the molecule containing I that they are designed to test. So test kits reading should not be compared to ICP-OES.

BRS you may want to remove reviews that compare Triton test results to molecular test kit results. The 2 types of testing measure different things."

So is this basically just a matter of ICP reporting total iodine while hobby kits only measure one or two specific forms? The Red Sea kit on BRS doesn't even specify as far as I can see, looks like Salifert does.