3 ms·
If anyone cares to increase the resolution of a rapid test, I recommend using Fiji and treating the image of the test strip like an antibody blot image and run
by pama 3y ago
If anyone cares to increase the resolution of a rapid test, I recommend using Fiji and treating the image of the test strip like an antibody blot image and run the quantitative gel analysis (easy to find info on the web). The software integrates the signal along the direction of the strip, correcting for lighting conditions, and the resulting single curve can clarify a little bump well before (or after) the viral load is strong enough to be visible by eye. I haven’t had covid myself yet (at least not tested positive) but I used this technique successfully multiple times on relatives and friends even on celebrity covid test pictures.
- greazy 3y ago> run the quantitative gel analysis (easy to find info on the web) You lost me at this bit. Use Fiji to run the gel analysis or is this via ImageJ or something else?
- pama 3y agoApologies for the confusion. Yes, run the gel analysis using Fiji. (Fiji is is a package distribution of ImageJ.)
- someperson 3y agoInteresting. That processing pipeline should be implemented into a phone app
- kbelder 3y agoIsn't this just shifting the balance between false negatives and false positives?
- pama 3y agoTo some degree yes, but these particular rehydrated antibody tests leave very little room for genuine false positives. If something binds on the test part of the strip it has the right epitope and there are no other viruses yet with these patterns. The integration of the signal across the width of the strip increases the signal to noise by a lot. I could detect decaying curves post infection at 50 times lower viral load than naked eye and confirmed by PCR. As the infection progressed, at some point the signal merges with the noise while PCR can still detect it, but the window is much smaller than with naked eye.