Smartphone Networked Optical Workflow
How SNOW reads a test
A lateral-flow test develops colored bands or spots on a membrane, and when one is faint, reading it by eye is a judgment call. A technician and a clinician can make that call differently, and neither can go back and check it. SNOW measures each one from a photo.
Three results on a canine tick-borne panel
The drawings differ only in which spots developed and how strongly.
Only the control spot developed, so the test is valid and the result is negative.
The control is strong, and one test spot is barely visible against the membrane.
Two test spots developed, but the control did not, and without a working control neither spot can be interpreted.
Bands, spots and dipstick pads
Most lateral-flow tests develop bands across a strip, and some, like the canine tick-borne panels, develop spots on a keyhole-shaped membrane. SNOW reads both. It finds each zone wherever the cassette puts it, so the same software can serve a whole test menu.
The third format is the urine dipstick, a strip with a row of square pads that each change color. Our dipstick reader corrects the photo's color against the strip's own blank pad, then matches each pad to the manufacturer's color chart.
Matching color is a different problem from measuring a band, and one a phone camera is well suited to. Doing it reliably takes its own development work, so the dipstick reader is a separate project, now in clinical evaluation.
Not yet validated
In clinical evaluation
Four separate layers
Each layer can fail without taking the others down, so SNOW can be useful in a clinic long before the disease map has enough coverage to show a trend.
Photograph
A technician takes the photo on an ordinary phone, and the software guides the shot and identifies the cassette type. It rejects an unreadable photo at the bench, while the test is still in hand.
Read
SNOW finds the result window and straightens it, so a cassette photographed at an angle or upside down reads the same as one squared up under good light. Because each band or spot is then measured as optical density against the membrane immediately around it, clinic fluorescents and daylight give the same reading.
Record
Each record holds the photo, the software's own measurement and the clinician's call. If the call and the measurement disagree, both stay on file.
Map
This layer shows SNOW's own readings, each with its date and, when the owner's ZIP code is on file, its county, on a county map beside public surveillance data, such as human case reports from the CDC and state health departments. Each source stays on its own layer. SNOW's layer works in rates rather than raw counts, so a practice that tests more does not look like one seeing more disease, and it leaves blank any area with too few tests to say anything.
Measuring with an ordinary phone camera
A phone is not a controlled instrument. Cameras differ from one phone model to the next, their image processing is proprietary and changes without notice, and clinic lighting varies. Any approach that depends on absolute brightness fails the first time someone stands near a window.
So SNOW leaves absolute brightness out of it and measures the difference between each zone and the membrane immediately beside it. Bright light makes the band and the membrane both lighter and dim light makes them both darker, but in either light the difference between them is still there.
Invalid tests and unreadable photos
If the control did not develop, SNOW reports the test as invalid, not as negative. It also names problems with the photo, such as glare across a zone, a window cut off at the edge, or a cassette too small in the frame. For each one, it tells the technician what to do.
When a clinician disagrees with a reading
The original measurement is kept next to the clinician’s correction. Those corrections are the most useful records the software holds, and we use them to check it and improve it.
Current status
SNOW is in development and is not a cleared or approved diagnostic device. We calibrate its readings of the canine tick-borne and heartworm panel against our internal sets of labeled reference photographs, and we keep refining it as more readings come in. The feline panel is not yet validated. The urine dipstick reader is in clinical evaluation, where its readings are compared pad by pad with the clinician’s own read of the same strip. Every reading is presented for a clinician to confirm and is no substitute for their judgment.