DAVID GREENE, HOST:
Now let's follow up on the story of software that lets parents check their children's eyes. It's one more use for your smartphone, searching for symptoms of a dangerous illness. NPR's Joe Palca's been talking with the man who developed this smartphone app. His name is Bryan Shaw, and here is the story of how the software became a reality.
JOE PALCA, BYLINE: Here's what got Bryan Shaw going on his software plans. His son Noah had to have his right eye removed as part of his treatment for a rare form of cancer called retinoblastoma. Shaw saw signs of his son's cancer long before doctors did - a white reflection coming from Noah's eyes in flash pictures. The flash was reflecting off tumors at the back of Noah's eyes. Shaw started seeing that white eye in pictures taken with his digital camera when Noah was just 12 days old. Doctors didn't catch the disease until months later.
BRYAN SHAW: If I would have had some software telling me, hey, go get this checked out, that would've sped up my son's diagnosis, and the tumors would've been just a little bit smaller when we got to them, there might have been fewer.
PALCA: And maybe Noah wouldn't have lost his eye. Since the software didn't exist, Shaw decided he'd create it. Now, Shaw isn't a computer programmer; he's a chemist at Baylor University in Waco, Texas, so he looked for someone in computer science at Baylor who could help him. He found Greg Hamerly. Hamerly and some of his grad students wrote a program that uses something called machine learning to spot a white reflection in photos.
GREG HAMERLY: Machine learning is about training the computer to do things by example.
PALCA: Using pictures of Shaw's son Noah and others that Shaw collected from parents who heard his story on MORNING EDITION, they trained the computer to recognize white-eye. They also trained the computer what normal eyes look like. That was easier.
HAMERLY: We just go out and gather images on the web that are public domain, and we can get thousands - millions of images of eyes.
PALCA: They've now turned a version of their white-eye detection software into a free iPhone app. I caught up with Shaw last week at a conference. He pulled out his phone and showed me how the app works.
SHAW: So the easiest thing to do is to just search your photos for the white eye. And you just hit the button search photos and boom - it starts searching all your photos on your device, and it'll let you know if it sees anything.
PALCA: You said it'll search what's on there, but it will also - you can use it - like, you could point at me and look to see if I had any...
SHAW: Yeah, you can also hit what we call screening mode.
PALCA: In screening mode, the phone shines a light into your eyes and the phone's camera looks for a white reflection. The app says my eyes are normal. And getting normal right is critically important. If the app makes a mistake and says there's an example of white-eye in a small child when there really isn't, that could easily freak out anxious parents.
JANE EDMOND: That's what I would be worried about, that there'll be over-referral and panic.
PALCA: Jane Edmond is a pediatric ophthalmologist at Texas Children's Hospital. At this point, it's not clear just how many mistakes the new app will make. But if the app does routinely get it right, Edmond says it could be a useful tool. Edmond says the real value of the new app may be in countries with lots of cell phones but not a lot of hospitals or clinics.
EDMOND: Think about places that have no doctor. Could this be - it's not a substitute for a doctor, but is it better than nothing? Well, yeah.
PALCA: And Bryan Shaw is hoping the new app will be a lot better than nothing. Joe Palca, NPR News.
STEVE INSKEEP, HOST:
That report was part of the series Joe's Big Idea.
NPR transcripts are created on a rush deadline by Verb8tm, Inc., an NPR contractor, and produced using a proprietary transcription process developed with NPR. This text may not be in its final form and may be updated or revised in the future. Accuracy and availability may vary. The authoritative record of NPR’s programming is the audio record.