Start-Ups

With Partpic's visual recognition tech, you'll never buy the wrong part again

Atlanta-based Partpic lets users search and accurately identify product parts using only a photo. Here's how they've taken the startup world by storm.

partpict.jpg
The Partpic team is lead by co-founders Jewel Burks (center) and Jason Crain (second from left).
Image: Loyall Hart/Partpic

If you've ever worked on a home improvement project, or even put a bookshelf together from your favorite big-box store, you know that it can be difficult to keep track of all the tiny nails, screws, and other parts you need to complete the project.

The peak of frustration occurs when you need to replace a piece or buy an identical one, but you don't know the part number or even what it's called. You call the home improvement store, but there's only so much you can tell them over the phone to describe what you need. If you're brave enough to make the trek to the store, you find yourself sifting through million different bins trying to find what you're looking for.

[Cue U2's "I Still Haven't Found What I'm Looking For"]

These are the problems that Atlanta-based startup Partpic is trying to solve. The company has developed a proprietary visual recognition technology that allows users to search for a part by simply taking a photo.

Once the image is taken through the app or uploaded to it, Partpic begins searching its growing database for identical parts. The technology can recognize a part down to its exact number, not just a category. It won't just tell you the part is a screw — it will tell you what type of screw it is and what distributor carries it.

A process called feature extraction pulls out the different features in a given part. The algorithm measures the part itself, its head type, finish, and thread count to determine what it is. Right now, they are focused on fasteners, which are things like screws, bolts, and rivets.

"All these are characteristics that make a given fastener unique," co-founder and CEO Jewel Burks said.

partpic2.jpg
The Partpic demo app.
Image: Partpic
The company will license the technology to industrial distributors, retailers, and suppliers so that they can use Partpic technology on their own websites or mobile apps. The technology can even be integrated into a kiosk system so folks could bring their parts in, scan them in-store, and be directed to the correct aisle and location.

The idea for Partpic was hatched roughly two years ago when Burks was working at industrial distributor McMaster-Carr. As a sales supervisor, she often dealt with upset customers who had been sent the wrong parts by mistake. She had also run into the same problem personally, trying to locate something for her grandfather.

"My grandfather was looking for a tractor part and I couldn't find it," Burks said. "He had called me to help him find it, because he knew I was at a parts company, and I was having the hardest time trying to find this part."

She called up Jason Crain, whom she had previously worked with at Google. Crain was working for music recognition app Shazam at the time, and as the two were brainstorming the problem, the idea of image recognition for parts came about.

Starting out, the co-founders thought they would build the technology using an existing API, but quickly learned that plan would not work. There was nothing on the market robust enough for what they wanted to do. So, they had to build it from scratch.

After reading that Georgia Tech had a great program for digital signal processing, which is the foundation of Partpic's technology, Burks started to get involved in the community there. The campus was around the corner from her house, and she met some people in the doctoral program who were interested in the technology. Those people became the technical team, and they have been building out the product for about a year.

Burks and Crain invested their own resources in the company early on and began competing in startup pitch competitions around the country, winning many of those they entered. So far, they have won or placed in the following competitions:

Additionally, the team will be pitching July 4th weekend as a finalist at PowerMoves NOLA.

So far, the team had been funding Partpic from the competitions, which often come with a monetary award. But, all of those competitions netted them more than just money and bragging rights. They eventually led to their first formal round of funding: a $1.5 million Seed round.

"Raising the money allowed us to bring on all of our engineers full-time, and actually have the staff to accommodate all of the interest we have on the customer side," Burks said.

Angel investor Joanne Wilson, who invested in the Seed round, said that she heard about Burks and was intrigued so she asked another investor for an introduction. She was impressed with Burks personally, but also with what the team had built. She said they are filling a void.

"It will streamline the way that customers are able to locate parts vs having to pour through large tomes to find the right piece," Wilson said. "It will make take this business into the 21st century. It is a win for every vertical that uses parts, from cars to contractors to medical devices."

Currently, Partpic is working with two companies who are also investors. Burks said they have the companies' parts in the system, and they are about a month out from launching Partpic on their respective websites.

Partpic is laser-focused on fasteners, but they could expand to other industries in the future. Burks and Crain said they are interested in automotive, as there are a great deal of fasteners that are unique to automotive and they've gotten a lot of outreach from auto industry.

"That's probably going to be the next big frontier for us," Burks said.

Also see

About Conner Forrest

Conner Forrest is a Senior Editor for TechRepublic. He covers enterprise technology and is interested in the convergence of tech and culture.

Editor's Picks

Free Newsletters, In your Inbox