Take picture every 2 minutes until completion after sample inserted

[pld]

We'd like to derive information from the rate at which a test strip changes color to investigate if that contains

• information on parasite-load
• information on whether the test is broken, e.g. does an image after 4 minutes correlate better w/result of PCR than image after standard amount of time

We also want to do this to see if there's a normalized change in color/intensity/etc that is more predictive than absolute intensity, e.g. is intensity of 10 minute image - intensity of 6 minute image a better predictor of outcome?

[shivenbhatt1]

I'm starting to question the need for this and what it will prove

[pld]

Good question, I should have added more detail, and even w/that it still might not be worth doing, I updated the description to add:

We'd like to derive information from the rate at which a test strip changes color to investigate if that contains
* information on parasite-load
* information on whether the test is broken, e.g. does an image after 4 minutes correlate better w/result of PCR than image after standard amount of time

We also want to do this to see if there's a normalized change in color/intensity/etc that is more predictive than absolute intensity, e.g. is intensity of 10 minute image - intensity of 6 minute image a better predictor of outcome?

I also think there were other things we hoped to get out of interval pictures, we should ask Anu for his ideas before we make a decision.

The ML data hoarder in me would also want to do this under the principle that if there is not signal here the only way we can know that is if we collect the data.