캘리포니아에 위치한 벤처 기업 ‘스탠다드코그니션(Standard Cognition)’이 최근 $5.5 million(약 6조) 규모의 투자금을 유치한 것으로 밝혀졌다. 스탠다드코그니션은 아마존고(Amazon Go)가 갖춘 화상 인식, 보안 카메라 기능 등 동일 기술을 개발하는 기업이다.
스탠다드코그니션은 어플리케이션을 통해 개개인이 어떤 물건을 장바구니에 담는지 파악하고, 쇼핑이 종료되는 시점에 개별 계산서가 청구된다. 이때 모든 데이터는 익명으로 수집된다.
스탠더드코그니션의 공동 창립자 겸 COO인 마이클 수스월(Michael Suswal)은 “현재 플랫폼은 이미 완성되어 있으며, 계약 체결을 한 각 파트너에게 플랫폼을 공급하는 데 주력하고 있다”라며 “업체마다 다른 운영 방식에 맞춰 유연하게 대처할 수 있도록 노력하고 있다”라고 말했다.
아마존고와 동일한 기술을 제시하지만 데이터 수집 방식에 있어서는 차이를 둔다. 매장 내 단순히 카메라만 설치하는 것을 넘어, 데이터 수집에 필요한 기기들을 활용하고 고화질 비디오 외에도 다양한 정보를 제공할 수 있을 것으로 예상된다. 예를 들면 소규모 업체는 공간 확보로 더 많은 소비자를 수용할 수 있고, 제품 판매 현황을 쉽게 파악할 수 있기 때문에 이전보다 더 높은 효율을 낼 수 있다. 특히 아마존고와 같은 무인 매장이 늘어나고 있는 시점에서 매장 현황 데이터를 확보하는데 도움 되는 플랫폼이 될 수 있을 것이라 예상된다.
마이클 수스윌은 “사생활 보호를 위해 클라우드에 얼굴인식과 같은 데이터를 저장하지 않지만, 신규 업데이트와 매장 간 교차 분석을 위해 최소한의 정보만 남겨 놓는다”라며 “데이터는 수집 목적이 아닌 새로운 기능 추가를 위한 것”이라며 재차 강조했다.
한편, 머신러닝의 발전으로 소기업에서도 기술기반의 시스템을 도입하는 것이 이전 보다 수월해졌다. 이미 많은 회사에서 음성인식 기술인 사운드하운드(SoundHound)나 이미지 캡쳐, 프로세싱을 실행하는 머신비전 기술인 칼리파이(Clarifai)를 도입하는 추세다. 물론 자본 차이로 앞으로 업체 간 기술 점유 부분의 경쟁은 발생하겠지만 혹자들은 스탠다드코그니션이 아마존고에 비견될 수 있을 만큼의 유의미한 규모의 매장 확보를 할 수 있을 것으로 기대하고 있다.
As Amazon looks to increasingly expand its cashier-less grocery stories — called Amazon Go – across different regions, there’s at least one startup hoping to end up everywhere else beyond Amazon’s empire.
Standard Cognition aims to help businesses create that kind of checkout experience based on machine vision, using image recognition to figure out that a specific person is picking up and walking out the door with a bag of Cheetos. The company said it’s raised an additional $5.5 million in a round in what the company is calling a seed round extension from CRV. The play here is, like many startups, to create something that a massive company is going after — like image recognition for cashier-less checkouts — for the long tail businesses rather than locking them into a single ecosystem.
Standard Cognition works with security cameras that have a bit more power than typical cameras to identify people that walk into a store. Those customers use an app, and the camera identifies everything they are carrying and bills them as they exit the store. The company has said it works to anonymize that data, so there isn’t any kind of product tracking that might chase you around the Internet that you might find on other platforms.
“The platform is built at this point – we are now focused on releasing the platform to each retail partner that signs on with us,” Michael Suswal, Co-founder and COO said. “Most of the surprises coming our way come from learning about how each retailer prefers to run their operations and store experiences. They are all a little different and require us to be flexible with how we deploy.”
It’s a toolkit that makes sense for both larger and smaller retailers, especially as the actual technology to install cameras or other devices that can get high-quality video or have more processing power goes down over time. Baking that into smaller retailers or mom-and-pop stores could help them get more foot traffic or make it easier to keep tabs on what kind of inventory is most popular or selling out more quickly. It offers an opportunity to have an added layer of data about how their store works, which could be increasingly important over time as something like Amazon looks to start taking over the grocery experience with stores like Amazon Go or its massive acquisition of Whole Foods.
“While we save no personal data in the cloud, and the system is built for privacy (no facial recognition among other safety features that come with being a non-cloud solution), we do use the internet for a couple of things,” Suswal said. “One of those things is to update our models and push them fleet wide. This is not a data push. It is light and allows us to make updates to models and add new features. We refer to it as the Tesla model, inspired by the way a driver can have a new feature when they wake up in the morning. We are also able to offer cross-store analytics to the retailer using the cloud, but no personal data is ever stored there.”
It’s thanks to advances in machine learning — and the frameworks and hardware that support it — that have made this kind of technology easier to build for smaller companies. Already there are other companies that look to be third-party providers for popular applications like voice recognition (think SoundHound) or machine vision (think Clarifai). All of those aim to be an option outside of whatever options larger companies might have like Alexa. It also means there is probably going to be a land grab and that there will be other interpretations of what the cashier-less checkout experience looks like, but Standard Cognition is hoping it’ll be able to get into enough stores to be an actual challenger to Amazon Go.
[fbcomments url=”http://ec2-13-125-22-250.ap-northeast-2.compute.amazonaws.com/2018/07/18/interpress-standardcognition/” width=”100%” count=”off” num=”5″ countmsg=”wonderful comments!”] ]
