Why AR for Enterprise » Technologies for Enterprise AR

Image Recognition

f-imagerec_fl_090914-8

Nearly a decade ago, when the first camera phones (mobile phones with built-in cameras) were released and began to reach the mass market, new services and software began to emerge. Photo blogging platforms proliferated, though the user still needed to download images to their computers before publishing them.

At the time, there were several options for video calling using a camera phone, although such phones could not yet be called smartphones. Such video software often used the same channel as the voice service (circuit-switched network) to establish the connection and transmit six video frames per second. Where 3G was deployed, network coverage was spotty and bandwidth insufficient. Then there was the problem of whom the user was going to call when, prior to 2009, camera phones only constituted less than 20% of phone devices.

Quick Response Bar Code Scanning

A promising technology was Quick Response (QR) bar code scanning. A photo taken of a Quick Response code (a scan) was analyzed and a small sequence of encoded digits was transmitted to a server in the network (no one was thinking of cloud services in those days). In Japan and Korea, QR codes became popular relatively quickly. QR code technology became an ISO international standard (ISO/IEC18004) in 2000, and a GS1 standard for mobile phones in 2011. More information can be found here.

Many companies published QR barcode scanning applications. Some companies also invented proprietary visual bar code systems so they could lock in vendors who signed up with their platform. While QR codes are frequently seen on products today, especially for application downloads, they did not gain global popularity for simple tasks and, despite being piloted and promoted in enterprise for inventory management, these did not rival or fully displace the use of RFID.

kooaba

Companies that developed image recognition had to find other uses for their technologies. For example, kooaba, a spin off of the Computer Vision Lab of the ETH Zurich and founded in 2007 by Till Quick and Herbert Bay, was an early pioneer in the field of image recognition. They found a niche in the area of interactive print. Image-based advertisements in magazines, newspapers and posters are detected and recognized, and the user is sent to a web page configured for a particular promotion. Using the management platform commercialized since 2012 by a kooaba sister company, Shortcut Media, the offer can be customized for a geographic region or other constraints, to best meet the interests of the consumer. In early 2014, kooaba was acquired by Qualcomm Connected Experiences, Inc.

Image recognition providers

Who does that leave?

TechNavio, a London-based market research publisher, recently released a new study focusing on the business of image recognition. The report assesses the drivers and obstacles to the growth of the image recognition market. It concludes that over the next few years, image recognition’s value for advanced screening and security enforcement (primarily in government sectors) will be a key growth driver. Lower-than-desired accuracy (the shortfalls of existing technologies to correctly recognize images and objects) remains an important barrier to growth.

The report profiles 12 image recognition software providers, including three with an emphasis on Augmented Reality.
Catchoom
Honeywell International Inc.
iTraff Technology Sp. z o.o.
Ltutech
Blippar
Hitachi
NEC
Panasonic
Qualcomm
Sharp Vision Software
Toshiba
Wikitude

TechNavio Global Image Recognition Landscape, July 30, 2014.
TechNavio Global Image Recognition Landscape, July 30, 2014.

Enterprise AR developers should monitor the evolution of image recognition for both 2D and 3D objects, as this can positively influence the robustness of AR solutions when seeking to identify targets in the business environment.