Archive for the ‘Artificial Intelligence’ Category

  • Jan
  • 16

The Future of Mobile Search & Augmented Reality

Posted in Artificial Intelligence, Augmented Reality, Image Search, Interactive Marketing, Mobile Marketing

What are your predictions for image recognition and augmented reality in 2010? In the last few months of 2009, we saw a meteoric rise in the number of developers, brand advertisers, media publishers, retailers and technology companies engaging with augmented reality, and in the case of Google, image recognition-based augmented reality; i.e., Google Goggles visual and location aware search through the mobile camera.  What types of AR do you think we’ll see in the coming months?

Twitter has helped facilitate many great discussions on AR, mobile image recognition and futuristic ideas that are coming to life in the present.  People like Chris Grayson of GigantiCo, Richard MacManus and Marshall Kirkpatrick of ReadWriteWeb, Rachael King of BusinessWeek.com, Tim O’Reilly of O’Reilly Media (see his Web Squared video with John Battelle for insight into image recognition and the future of the Mobile Web) and many others, are evangelizing augmented reality applications and core technologies.

As a provider of image recognition capabilities, Pongr is interested in helping you, your company, and your customers continue to innovate in this exciting space.  After all, the global heartbeat of the mobile Internet is connected to the billions of pictures generated by mobile consumers everywhere.  Pictures and videos are the lifeblood of how we share experiences and make connections beyond what can be said in 140 characters or less!

While we can’t predict the entire future of AR, we are certain of one thing: image recognition is a fundamental technology that will improve the way the world searches for, and interacts with people, places and things. We’d love to hear your thoughts on what you think will happen next.  Here are a few areas that we expect will continue to rage in 2010:

  • Image recognition as a core mobile “search” tool
  • Augmented reality that includes computer vision vs. GPS/LBS only AR Lite
  • Retail uses of AR in the form of “shopper technology” to drive traffic and sales
  • Gimmicky one-trick ponies vs. sustainable AR applications
  • Mingling of relationships between media, mobile and advertising companies


Read More... Write Comments (2)
  • Nov
  • 15

Going Rogue: A Mobile Marketer’s Guide to Augmented Reality Campaign Planning

Posted in Artificial Intelligence, Image Search, Interactive Marketing, Mobile Marketing, Social Media

The Economist, BusinessWeek, Harvard Business Publishing, and hundreds of others in the mainstream media are talking about augmented reality and the implications it will have on ad campaigns.  In fact, most of the attention being paid to AR Lite marketing is not only about the future of mobile marketing, but what’s being rolled out now!  Shopper technology is everywhere, Internet titans are staking out new territories, and the couplings between social media and mobilized ad campaigns are growing tighter.

But here’s the thing — we’re only seeing the tip of the iceberg in terms of the effectiveness of actually influencing consumer shopping decisions via mobile banner ads.  What lurks beneath the surface is huge, and the implications of what happens next are significant for the business of retail, advertising, and even wireless carriers.

As Adweek’s Brian Morrissey writes in a recent article on Google’s acquisition of AdMob:

The big purchase is a clear endorsement that Google sees enormous potential in a market that’s still small. Despite the never-ending prediction of “the year of mobile,” the market has grown by fits and starts. According to eMarketer, advertisers spent just $320 million on mobile ads last year, a figure that’s forecast to increase to $416 million this year and reach $1.6 billion in 2013.

Here are some practical tips on how to include augmented reality in your mobile marketing campaigns.

Rule #1 – It’s all about the opt-in experience!

Your best users and biggest fans need to be empowered to participate in a cool, mobile experience when it’s convenient for them.  That means you should think carefully about whether you want to associate your brand with “pushing” a mobile offer when it’s not directly solicited or guaranteed to be relevant.  Instead, you should tell a story; a story that involves a mobile component, is tied into your overall marketing activities, and includes some specific special offers and social opportunities for your fans to engage with your brand on their terms.

To be clear, when we say special offers, we don’t necessarily mean just coupons.  Mobile coupons could be a sensible way to approach your objectives, but there’s a lot you can do with the post-click in mobile.  Sign-ups for insider information, mobile videos, mini-sites, Twitter and Facebook pages, product info, and game tips are a few thoughts. For some additional suggestions on what to “offer” mobile users see Pongr’s page on mobile marketing “ad-ventures” for brand image recognition.

Rule #2 – Be inclusive. Use MMS for maximum reach and supplement with smartphone experiences.

In our humble opinion, you need to be as inclusive as possible with your special, opt-in mobile offers. That means you need to acknowledge that there’s more to a successful mobile marketing campaign than just an iPhone or Android app.  Specifically, we recommend factoring MMS into your campaign plans.  It goes without saying that brands should be building, maintaining, and enhancing applications on smartphones, but to reach the masses you must do something more pervasive.

SMS is great, but it doesn’t give you the ability to be as interactive as you could.  MMS is where it’s at when it comes to getting the base of the pyramid to opt-in to your campaigns and provide unique insights about themselves.  If you want to collect user photos, encourage sharing on social networks, and build a memorable experience that takes advantage of the ubiquity of picture sharing on the Web, you need to have an MMS plan for all those feature phone users.  The best plan in the world won’t move the needle if it doesn’t have the reach.

“You’ve got to be able to run on all phones, not just smartphones.”Alistair Goodman, Placecast

Rule #3 – Be clever & take advantage of the latest technology, but keep it simple!

The third rule of getting augmented reality into your ad campaigns is to keep it simple. OK, nothing Earth-shattering here, but as we all know, the more complicated it gets, the less likely you are to have a successful outcome.   Worse, your target users might be left feeling dumb or betrayed by your brand because you didn’t include them, or set the bar too high with something too complicated.  We’ve written a number of posts and tweeted a lot about QR codes, so we won’t rehash it all now, but fundamentally, Pongr believes a pure-play image recognition solution is better for users, creative directors, brands, and pretty much everyone.  Exceptions include enterprise use-cases like airline ticketing or monopolized wireless carrier countries, like South Korea.  Besides, if you want to look magical, why not go for the best, keep it easy, and keep it omnipresent.

Augmented reality is not new to many of us in the artificial intelligence/machine learning community and there’s been plenty of moderately nifty consumer applications out for a number of years.  However, with major brands, publishers, media companies, and consumer technology providers embracing the underlying components of augmented reality, we’re all about to see some dramatic & widely used AR and AR Lite applications in mobile marketing.

Advanced image recognition technologies and MMS advertising solutions are here now.  If we’re even close to being correct, image based search and other AR technologies will soon influence how you effectively send & receive message to customers.

If you’d like help planning or executing campaigns using augmented reality, image recognition, MMS, or mobile-to-social brand interaction, give us a call or zing us on Twitter!  Pongr provides leading-edge image recognition technology and turn-key augmented reality marketing solutions.

If you’d like to share your two cents, we definitely want to hear it!



Read More... Write Comments
  • Oct
  • 24

Augmented Reality and AR Lite for Mobile Marketing

Posted in Artificial Intelligence, Image Search, Mobile Marketing

The “Augmented Reality” hype train continues to gain momentum, and like the hackneyed idiom goes, where there’s smoke, there’s fire.

Below are a few tips we think you should keep in mind when doing your research on mobile augmented reality apps, the design of your own mobile marketing initiatives, and, for you futurists, your predictions about the Mobile Internet & Smart Phone Tsunami . . .

  • Nearly everything you see today that is being called Augmented Reality is a mashup of 2 things: the GPS (or non-GPS location identifier like WiFi or cellular triangulation) on the phone and 3rd party data APIs (e.g., Wikipedia, Google Maps, Yelp, eBay).  From now on, we’re going to refer to this set of mobile applications as AR Lite apps.
  • The use of the camera (or open video stream) is currently mostly about style and interface.  Some mobile phones (iPhone 3GS and some Android handsets) are using compass capabilities to provide a general sense of perspective in the app – that’s what separates an LBS enabled mobile app from this newer set of AR Lite apps.
  • The compass is what tells the AR Lite application where the user is “looking” – not the camera.
  • If you want a granular sense of perspective via the camera, you should use image recognition.  Like the AR Lite apps, a full-blown image recognition enabled mobile app is not magic (although it is magical and full of science), so you should learn more about this topic on its own.  Pongr is happy to help.
  • There are major wireless network and video compression matters that need to be fully addressed/rolled-out before mobile application developers will be able to utilize an open streaming camera view for the purposes of search (whether general purpose or vertical specific).  From a practical perspective, it doesn’t make sense to rely on high-quality streaming video for search just yet.  It will in in a few years, though.
  • We are likely to see a bifurcation among the AR Lite apps and image recognition based applications before we see them blend into a fully-functional, mass marketed video type of search.  The blend will occur only after the major carriers (or even just one) rolls out 4G/LTE networks. Even then, it remains to be seen what the data capabilities will be.
  • It is possible, if not likely, that carriers will have a heavy hand in controlling this next generation of mobile search because the consequences to their network assets (and profitability) are significant.  We may see alternative data infrastructure providers/sources, but that’s another topic altogether.  And, of course, in the U.S. and other countries, the network neutrality issues will greatly shape what happens with mobile video based search.

Contact Pongr or @Jamie_Thompson for more information.



Read More... Write Comments
  • Sep
  • 02

An Insider's Look At Mobile Augmented Reality

Posted in Artificial Intelligence, Image Search, Mobile Marketing

Here’s the skinny on mobile augmented reality: the apps you’re seeing right now are only using half of what augmented reality will become.  In terms of layered information on top of a mobile camera view, location data is powering most of what has been labeled “augmented reality” apps by the press and many pundits.  Some degree of perspective is available from the compass in the new iPhone, and certainly other phones will be getting this additional data source to fuel their own versions of quasi augmented reality apps, but the meat and potatoes of the apps that encourage you to keep an open camera or video stream are not yet using any type of wide-scale image recognition.

At the moment, it’s easy to get confused with this and naturally assume that because you point your phone’s camera at an object and information pops up in the display, it must have something to do with the camera interpreting what it sees, right?  Not the case.  In fact, in some apps you can actually put your finger over your mobile camera and you’ll still see the same data.  Try it!  Of course, it won’t look as cool, but the information will remain unchanged.  Actually, if you do something like this while moving in an automobile at about 30 MPH, or faster, you’ll notice the degree of accuracy goes down dramatically.  Don’t try this if you’re the driver – just take our word for it!  This latency, or bottleneck, as you might think of it, has nothing to do with the lack of image recognition or the use of location data, but more to do with the multi-functional nature of the mobile phone and the wireless network itself.  In time, this will not be such a noticeable issue as data capacities grow faster and faster.

As you might imagine, the personal navigation device market and built-in auto nav systems are not likely to be replaced by mobile augmented reality apps anytime soon…  On the contrary, they are more likely to gain additional sensory input mechanisms via cameras connected as accessories streaming data to back-end, cloud-based image recognition systems. It might sound complicated, but it could actually be pretty simple. You probably won’t see anything like this for the upcoming holiday shopping season, but maybe the year after…  The current batch of data being served up by augmented reality apps is definitely relevant when you consider close proximity to the user, but it has little relevance in any sort of dynamic environment or brand-centric promotional world.

Now, thanks to the iPhone and others for allowing easy access to latitude/longitude, there has been a low barrier to entry around creating location-aware mobile applications.  The lat/lng combined with a plethora of 3rd party data APIs has been a real source of inspiration for mobile developers.  Likewise, with access to the camera, creative developers have been using the lens view (not just on iPhone, but other handset platforms like Android, BlackBerry, Nokia, etc.) to provide a sense of high-tech, Web-Meets-Physical augmented reality experience.  Expect to see a lot more of this soon.  Yet, far fewer developers have been tapping into the power of pure image recognition.  There’s a good reason why you’re not seeing the image recognition component yet; it’s not readily available for the masses and it’s hard to independently develop systems that are robust enough to work in the real-world.

2010 will bring an onslaught of heretofore unknown application developers, brand marketers (e.g., emerging advertising agencies and media integrators), new analytic tools and vertical mobile search engines (special purpose apps that cross more than one mobile platform and likely leverage unified mobile browser standards) to market in a powerful way.  Pongr’s own research and development in and around the mobile marketing industry has shown a tremendous ecosystem of media companies, new and old, helping to pave the way for some remarkable new mobile brand-to-consumer solutions.  All will be striving to use a few core pieces of augmented reality technology; image recognition and GPS (or other non-GPS location information like triangulation, WiFi, and even buoy tethered mesh networks).  The above will be accelerated or hindered against the backdrop of the global economy, but will grow significantly in terms of actual dollars spent on mobile originated image recognition and location awareness capabilities.  We base our estimates on our own internally collected data points as well as external indicators showing significant demand for core mobile augmented reality technologies.

The usual tech titans, global marketing agencies, brands themselves, and wireless infrastructure parties will be dipping in and out of the augmented reality waters too, but as is always the case with cutting-edge technology, expect to see the most disruptive advances in platform-oriented augmented reality from startups.

There’s also going to be some opportunities for a few older companies in the GPS space to reinvent themselves, or at least to bring a few new mobile products to market, moving further down the value-chain closer to the end consumer.  This could actually breath new life into some of the behind-the-scenes types of companies looking to build livelier consumer brands out of their existing base of users – users that have been faithful and loyal “customers” for years.  You can put two and two together here and say that existing brands that quickly figure out an executable strategy around augmented reality will have an advantage as the Internet shifts from being personal computer centric to being mobile and always-on, always-available.

OK, you want more? So this morning I was listening to Tim O’Reilly on a FutureTense podcast discussing his new concept “Web Squared” and how image recognition via mobile phone cameras will be a major factor in adding additional information about the world to the Web.  More specifically, mobile cameras are the sensors that augmented reality requires, with or without GPS, to truly unlock the potential of integrating social networks, real stuff in the physical world, and the vastness of the Web as it already exists today. Like Facebook, MySpace, Wikipedia, and even Google Maps, the power of augmented reality is not in the technology for the sake of technology, but the data that it provides at a moments notice.  Unlocking location awareness capabilities on the iPhone helped Apple go from 0 to 100 MPH in the mobile phone industry, and augmented reality becomes AUGMENTED REALITY when, for the masses, it is unlocked and fully capable of leveraging the real world mobile sensors we have all practically grown as a wireless data bodily appendage.  Weird, but true. Some people even have multiple mobile devices they use regularly. Think Edward Mobilehands.

While the words “augmented reality” are still a little squishy today, meaning many things to different people, the sticky part of “augmented” will occur when mobile users around the world have the ability to share and submit information they collect from their mobile camera sensors into a fully image searchable database.  Google, Microsoft, Yahoo and others have automated web crawlers, but the next great search engines will leverage mobile sensors, as well as the users for whom information is a two-way street of giving and getting – be it in a purely social network setting, a wikipedia-esque capacity, or a brand oriented conversation amongst consumers.  Some of the forthcoming information silos will emerge as better than others and certain types will lend themselves more readily to business value, but the commonality will involve mobile phone cameras, faster wireless networks, image search and location awareness.

To discuss in more detail or share your own thoughts on what’s going on with augmented reality, find me on Twitter at @Jamie_Thompson and @PongrMedia.



Read More... Write Comments
  • Jul
  • 21

Bridging The Gap Between TV, Print, In-Store, Mobile & Internet!

Posted in Artificial Intelligence, Image Search, Interactive Marketing, Mobile Marketing

Pongr is very excited to let you know that we have successfully linked television commercials to our mobile visual search system!  As many of you know, we have been re-focusing our efforts on building the most accurate and robust visual search system to allow mobile users to submit pictures.  Most people generally assume that this is limited to static images, however we’ve been conducting a number of internal experiments and customer proofs of concept around the notion of linking video into the system.  This morning, while waiting to join the daily development scrum call, a TV ad on CNBC caught my attention and I decided to give it a try.  Pongr matched the shot originating from my BlackBerry and returned the search results from Pongr’s automated mobile delivery service.  The entire process happened in the span of about 15 seconds on a normal mobile carrier connection (not WiFi, and with a bad signal at that).

While our visual search system still requires more data to make it as useful as possible for people, we are pleased that the application to TV is working as well as the application to print, billboard, in-store, etc.  We have an even more robust system still under development,  which will provide additional forgiveness to less than optimal user generated images/video, and can’t wait to give that a shot (on lesser quality TVs – not that this will be an issue for most users as time marches on in HDTV world).

The matching ad of the television search shot is here: Pongr Share page.



Read More... Write Comments