QromaTag is about photos and stories, and here’s mine. There were two events in my life that were key to building QromaTag. The first was losing my father when I was 28. I inherited his collection of photos and so many of them, especially the ones when he was young and traveling around the world, were a mystery to me. Many were striking in a way that made me wonder what they were all about. What was the story there? Where, when and why were those moments important? Without knowing the story, something about those photos always seemed missing. My second big epiphany was more ho hum. I worked at Apple when iPhoto was first released and soon I moved my growing photo library there. When iPhoto’s facial recognition feature was introduced, I was thrilled. Being able to find my photos based on people’s faces was amazing. I spent untold hours tinkering with this feature in hopes of tagging every face when disaster struck; database corruption. Without rhyme or reason; everything I worked for vanished. Gone were untold hours of effort, but I learned an important lesson about what kind of data survives and what doesn’t. From then on, whenever I spent time adding my stories to my photos, I used industry photo standard metadata written directly into each image, or I would never be sure if my work would be around for my children or their children.

Putting a Story inside a Photo

How many times have you looked at a photo and wondered what was going on? If it was a print, you might be tempted to flip it over to see if anything was written on the back. Digital images have the ability to store so much more that the back of any photo could ever hold but with acronyms like EXIF and IPTC, few get it’s full benefits.. For my 40,000 or so images, this wasn’t a problem. I wrote custom tools that helped me work around the fact that most photo organization programs didn’t ever actually write metadata back to the image. They merely stored all your photo information into their own database. I was always concerned with precision. It wasn’t ever good enough to say a photos was taken in San Jose; it had to be the exact GPS coordinates, with as precise a date as I could muster. I’ve always looked at my own photos thinking about how my children’s children’s children would interact with them. What questions would they ask? What would it mean to be able to impart much more than just a scribbled notation about something that was meaningful enough to photograph?

It’s always been technically possible to put a story inside a photo. That’s the beauty of standards. It just hasn’t been practical for most of us. This is why QromaTag came to be.

What is QromaTag?

QromaTag is an iOS application that makes it easy to put the most important parts of a story into any photo in a way that will survive for generations. Using two voice recognition systems that work in tandem, QromaTag creates industry standard photo metadata based on what you tell it about your photos. Using natural language, tell us the date, location and people that are in the photo and QromaTag takes care of all the technical details and embeds that information into the photo. Because these tags are industry standard, they will work on any platform, any device or any web service today, tomorrow and long into the future. With QromaTag, you can attach a description of of to 2,000 characters to any photo, ensuring that your future generation will experience your story in a way the ‘back of the photo’ never could.

How we built it

QromaTag will be the second product from San Jose, California based Qroma LLC. It shares more than a few genes with its sibling QromaScan, which is a book-like device that turns into a mobile scanning solution for iOS users. Both QromaTag and QromaScan share the ability to create industry standard metadata through the tandem use of two voice recognition engines. While QromaTag will see its first users in early 2017, it has benefitted from a year of upgrades, bug fixes and new features as QromaScan moved from version 1.0 to 2.0.

Challenges we ran into

By far, our biggest challenge was related to voice recognition. For detecting dates and locations, there were a number of ‘off the shelf’ web services that we could choose from that were reasonably accurate. Where they fell down significantly was recognizing proper names. These ‘online’ recognition systems, which work very much like Siri, simply didn’t contain phonetic transliterations for anything more than an handful of common names. We nearly gave up completely before we discovered something we had almost dismissed. It was a ‘local’ or ‘offline’ voice recognition system that was based on some excellent open source work from Carnegie Mellon University. It allowed us to dynamically generate phonetic sentences on a device by device basis and keep that information local to just that user’s phone. What this meant was that we could use the user’s Contacts or a GEDCOM file as the basis for generating a custom voice recognition vocabulary that could recognize proper names with a very high success rate. Over the past few months, we have worked on perfecting this ‘tandem’ of recognition engines in a way that helps us capture the most important details of a photo with the tool most of our users know best; their own voice.

Accomplishments we are proud of

I started with a simple idea about wanting to help people protect their memories and stories in a way that would survive to benefit their heirs. Slowly the idea evolved until it had a name. It’s a twist on the Greek word for color (chroma) and giving it a unique name helped give it life. Over the past 18 months, we have sold QromaScan Lightboxes in 43 countries and this word has become a part of other people’s lives. More importantly, our app analytics give us an idea of how many images were scanned and tagged during that time frame. Knowing that we played a part in bringing these stories and images into the digital age in a time when the use of physical scanners is declining is quite rewarding, and I think QromaTag can do even more.

What we learned

Loads and loads. I’ve had a nice career developing software for many well known companies, and things that I have worked on are in use by millions of people around the world. Building your own idea from the ground up has its challenges, rewards and learning opportunities. Perhaps the biggest lesson I have learned is that there is a delta between how you perceive something is going to be used, and how it actually is by people that were not involved in its design. Patience, customer input and iteration have helped us close this gap .

What's next for QromaTag

We are actively beta testing QromaTag now and working on developing the support material that will help users get the most out of it. We expect to launch worldwide on February 8th, 2017. We are also expecting to launch an Android version some time in 2017.

Get QromaTag

QromaTag is currently in beta testing, and we expect to launch it at RootsTech 2017. TestFlight versions are available for testing and judging purposes. You can download a before and after of the image that was feature in our demo video here. The zipped file includes the original untagged photo that was downloaded from Facebook, and the version that was tagged using QromaTag.

Share this project: