Snap a picture of any object with the CamFind app, and you’ll get the details on what the item is and where to buy it. The app, developed by Image Searcher, Inc., is “the only successful visual search engine worldwide,” and also translates into several languages. A breath-stealing car speeding past you, an eye-catching cute purse swinging down the sidewalk, or perhaps just a cup of coffee: Rather than googling the details on each, CamFind, available on both iOS and Android, allows you to skip the typing and go more directly to the results.
Image Searcher, Inc. is now fundraising on AngelList, with a goal of $1 million as convertible debt and a 10 percent discount in their current round. The company also had three previous rounds of funding: $650,000, which just closed Feb. 10, 2015 with a 30 percent discount; $400,000, which closed Dec. 19, 2014 with a 15 percent discount; and $322,000, which closed Oct. 15, with a 15 percent discount. The company also had previous funding of $4,812,000 before the current round. The interest rate is 4 percent per annum.
“It isn’t perfect, but the newcomer [CamFind] consistently outperforms [Googles’] when the two are compared side-by-side over the same item,” declared the website Cult of Mac. Since launching in April 2013, CamFind has processed over 13 million visual searches, according to its AngelList campaign page—and is “the only successful visual search engine worldwide.” The page also states that “all major media claimed that the app is far superior to Google Goggles.”
Here’s how the app works, according to its AngelList campaign page:
Human taggers log in to tag the images taken by iOS users of our app (so images which our algos can’t recognize). We tested over 7,000 people for the tagging position and hired less than 5% of them. We have a very sophisticated image distribution platform that took 1.5 years to build all in order to process images in near real time with humans. Whenever computer vision doesn’t get a good confidence index/coefficient, then the image goes to the crowd. So basically the crowdsourcing is used for the difficult 3D images that can’t be matched to anything in any existing database and can’t be handled by our deep learning neural networks.
Long term, the company plans to gain a significant market share in the $300B+ search market by outshine its few competitors in the visual search field. It also seeks to be “the leading provider of image recognition to corporations and retailers and therefore targeting the $25.65B image recognition market,” according to its AngelList profile.
Have a crowdfunding offering you'd like to share? Submit an offering for consideration using our Submit a Tip form and we may share it on our site!