The Augmented Reality Hype Cycle

Update: This article now also in Dutch. And in Catalan.

This year Augmented Reality (AR) is out of the box. It has escaped from the universities and is spreading like wild fire. It’s being written about more and more, especially by marketing people, who go for hypes first. This last month I’ve personally heard about at least 4 advertising agencies working on it. On Twitter its mentioned at least 5 times an hour. What types of Augmented Reality are there and where do we stand?

To frame this and differentiate the levels of AR I made the Augmented Reality Hype Cycle. It is inspired by Robert Rice’s article “Is it too early for Augmented Reality?” and expands on the AR Levels which he describes in “Augmented Vision and the Decade of Ubiquity “. The Gartner Hype Cycle is a great method to plot new technologies and applications.

In the Hype Cycle I focus on technology that is already out there and which application is aimed at the individual. Not Industrial. Also I focus on mobile when ever possible.


There are 4 Levels of AR (the larger grey titles):

  • Level 0 - Physical World Hyper Linking
  • Level 1 - Marker Based AR
  • Level 2 - Markerless AR
  • Level 3 - Augmented Vision

LEVEL 0 - Physical World Hyper Linking

This is the oldest form and an addition to Rices’s 3 initial Levels of AR. It is a way to link the physical world to the virtual world. It starts with the 1D (UPC) bar code that’s on every product you have. It’s the identifier as registered in the database. The same can be said for 2D codes of which QR-codes are best known. Last part of this group is 2D image recognition. Most used mobile application is the recognition of a company logo or film poster which will then redirect you to a site for more information.

Being the oldest, this form of AR is the most developed. Its also the simplest and doesn’t involve real time rendering and display of graphics. Some may not even call it AR. General adoption of this technology will happen within a year. Most Nokia phones are already outfitted with the software for 1D and 2D bar code scanning. 2D scanning is stuck on the slope of enlightenment and will be surpassed by 2D image recognition in the coming year.

Examples of 1D & 2D Barcodes:

Examples 2D Image Recognition

LEVEL 1 - Marker Based AR

This is where the hype is now. Augmented reality based on markers. Its the first step for real AR because it enables real time processing of ‘reality’ through the recognition of markers and subsequently real time rendering and display of graphics on top of this reality.

The first type is the 2D Marker AR which is PC and web cam based. Almost all applications work use an Internet connection to retrieve more information. The marker is the black and white square image you print out and hold in front of your web cam to see a 3D animation.

This is what most people are doing. There are also free toolkits to develop applications and now also Flash based which explains a lot of the popularity.

Next step up is 2D Marker AR with a mobile device. This is tricky-er because it involves a lot or real time processing and a very capable phone. For instance it is not possible on the iPhone because the OS does not allow for real time video processing. You also can’t record video with it as yet. It is possible with a hacked iPhone though.

2D Marker AR is most seen on Windows Mobile devices surprisingly enough. Coolest is 3D recognition, where the mobile recognizes a chair or another physical object and goes on to augmented it.

Level 1 - Marker Based AR is moving from the technology trigger to just before the actual peak of inflated expectations. For now it will not go away and in the coming year PC and Mobile 2D marker based AR will be the IN thing. Where the mobile version will be lagging in comparison to the PC based version due to the lack of computing power and other mobile issues. When the novelty is worn off the actual application value will prove to be thin. Its no real money maker except for marketing and PR applications. On the mobile it will last a bit longer when good games come out. The most promising, 3D recognition, is in the tail of Level 1 - Marker Based AR and is hardly a reality yet.

Examples of 2D Marker AR - PC & Web cam based:

Examples of 2D Marker AR - Mobile:

Examples 3D Object Recognition - Mobile:

LEVEL 2 - Markerless AR

Augmented Reality without markers is powerful. Pull out your mobile phone and experience an augmented reality. It can be that simple. Only G1 (the Android - Google phone from T-mobile) owners had the opportunity to experience this with Wikitude from Mobilizy and later ING Wegwijzer. Its based on GPS data and the compass. Because the phone knows where it is (the GPS) and in what direction you are looking (the compass) it can augment reality on the screen correctly.

Level 2 - Markerless AR is halfway the technology trigger and will be nearing the peak of inflated expectations in a year when more phones have the combination of GPS and compass. The Nokia N97 which will be out in the summer will have a compass and will trigger the first non Android Markerless AR applications. It is unknown if the new iPhone will have this capability. It would be HUGH if it did…

Examples of GPS - Compass based AR:

LEVEL 3 - Augmented Vision

This does not exist yet. As Rice says:

“We must break away from the monitor and display to lightweight transparent wearable displays (in an eyeglasses form factor). Once AR becomes AV, it is immersive. The whole experience immediately changes into something more relevant, contextual, and personal. This is radical and changes everything. As I have said before, this will be the next evolution in media. Print, Radio, Television, Internet, Augmented Reality (well, Vision). L3 must also be mobile massively multi-user, persistent, shared, dynamic, and ubiquitous.”

It will be more then 3 years before anything remotely capable will be available. Until then we will have to make do with research news like that of the University of Washington where they are working on displaying pixels in contact lenses .

In Conclusion

To me it’s daunting to see how far there is to go. Yet knowing where we are now, what is coming and how to frame it enables realistic planning and choices. I hope it helps you. I’ll be enjoying the ride and developing some cool AR initiatives with my SPRX partners.

Thanks to all AR bloggers and twitterers who directly and indirectly helped me put this together. Looking forward to see your feedback and more examples in the comments.

- April 2009

 Retweet This Post

Share and Enjoy:
  • del.icio.us
  • Digg
  • Technorati
  • Google
  • Facebook
  • Live
  • Ma.gnolia
  • Sphinn
  • Mixx

31 Comments »

  • huh? I don’t get your hype-cycle diagram. You’ve placed 1d barcodes at the very end, in the “slope of enlightenment” while augmented vision is right there at the beginning.

    Anyway, I’m not sure I would have divided the technologies into the four levels as you did. Nokia’s point and find, to me, is much more sophisticated and brings us closer to an augmented world than marker based AR. Marker based AR is essentially an easy to spot barcode plus (usually) a 3d rendered object. The image recognition part seems to me more challenging than the 3d rendering.
    Then, on level 2, you’ve placed gps+compass application, that don’t do neither image recognition nor 3d rendering. I think they are the easiest to implement, and although they are far more “AR” than those marker based application on your level 1, I can’t believe that at the end no image recognition will be involved (as a matter of fact, MIT’s sixth sense uses some image recognition - it must, for close by items).

    Sorry for being so critical. If I had to divide those technologies into level, I would have placed marker based AR on the first; GPS+compass apps and image recognition apps on the second; mobile GPS+compass+image recognition apps on the third; augmented vision on the fourth.

    Comment | April 27, 2009
  • [...] Op het SPRXblog meer uitleg en veel links naar voorbeelden. [...]

    Pingback | April 28, 2009
  • @rouli thanks for the comment, critical is good

    Nokia’s point and find is a form of marker technology with its image recognition on which they added category layer on top to sell it better. Its actually point and click…

    You are right that gps-compass AR doesn’t do 3D rendering. As yet. What I think is important about it is that the phone knows more about its context right away. You don’t need to do anythink, like …point and click.

    I also believe as you do, that this type combined with image recognition will be a great combination, especially for close up situations.

    Do you think RFID has a place in here? I think it might be a marker variant yet I dont know of any examples.

    Comment | April 28, 2009
  • So, in your opinion what divides technologies between the levels is the immediacy of the augmented layer? That is, since Nokia requires you to “click” in order to get the additional information, it should go the first level?

    If that’s the case, look for SREngine, http://artimes.rouli.net/2009/04/srengine-in-english.html, which combines immediate image recognition with GPS information.
    As for RFIDs, they can replace image recognition when it comes to close-range AR. They are indeed a marker variant in a sense, but a non-obtrusive one. Nokia’s next gen phones come with RFID readers and Alcatel-Lucent has an initiative to mass-produce RFID tags (http://www.touchatag.com/).

    Now, back to diagram, isn’t it reversed?

    Comment | April 29, 2009
  • Nice addition to http://yuri.typepad.com/yuri_blog/2008/04/mobile-augmente.html. Impressed :)

    Comment | April 29, 2009
  • What divides the different levels is how mature and capable the technology is and also how different.

    1 and 2D based apps, processes are very mature and capable. All cargo lives off of this. These technologies all are abour recognizing an image (1D, 2D or and actual image) and interpreting this into a link or an number which is linked to more info in a database or a web site. Nokia Point and click fits into this. Its is Level 0, the most mature, capable and there fore on the end of the graph. There is loads of money being made with this.

    Level 1 is AR based on markers. So again recognition of and image, either a 2D code or and actual image) plus the rendering of 2D or 3D graphics on top of the video input which subsequently is shown. I think this is not a mature or capable market yet. Its is new and has only gimmicky applications. People love it though. They will really like it until they find out that there is not to much you can do with it except for PR and marketing applications. That is why it placed it at the top of the hype cycle.

    Level 2 is compass and GPS based. With either 2D or 3D rendering. This is very new yet very promising because its easier for the individual (the app ‘knows’ alot with any markers. I do think these apps will mere with image recognition apps.

    etc etc

    mmm I dont think I understand well what your line up is based on. Mine is based on Capability and Maturity, the technology level of the sensor and the age of the technology. By sensor I mean how the app can sense its location and viewing direction. Its goes from simple marker to complex market to no marker. Also the diagram doesn’t not say any thing about the speed of development.

    I actually think that compass + gps AR will surpass marker AR and will merge ‘a little’.

    Hey, I think I know now what you based yr line up one. Its processor power. The more it uses the higher up you get?

    Looking forward to you reply.

    Comment | April 29, 2009
  • First, let me say that I finally understand the diagram, after googling other hype diagrams. Sorry for suggesting it may be wrong. BTW, it’s cool that yours is the third result when looking for ‘hype cycle diagram’.

    Yes, you can say that my line up is based on processor power. Actually, a better way to look at it, is the complexity of algorithms involved.

    I now think that in order to have a meaningful order on AR technologies, we have to examine them in two dimensions. The first is how the real life hyper-linking works. The second, is what kind of interface they use.

    Looking at the hyper-linking technologies:
    ———————————————-
    1d barcodes<
    2d barcodes and AR markers<
    2d image recognition (no design of special markers) ARToolKit, for example. Currently “2d barcodes and AR markers” is in mature state, and I think they are more well known than QR codes (in the western hemisphere).

    Looking at the interface:
    ————————-
    no interaction (GPS devices)<
    information comes in a new window (like in Point&Find)<
    text bubbles on top of the real video feed (Wikitude) <
    3d renders on top of the real video feed (ARToolKit)<
    3d renders that interact with the real world.

    The final step is so far off, I can’t think of an adequate example for it. The ideal is something like Denno Coil. Currently, the best example that I know of is WWF’s augmented reality (http://artimes.rouli.net/2009/04/wildlifes-fate-is-in-your-hands.html). Note that the 3d renders can be actually 2d text bubbles, as long as they are placed within our “3d world”.
    Currently, 2d text bubbles are mature enough, while 3d renders can still significantly improve.

    Hope I’ve made myself clear :). What do you think about this two dimensions perspective?

    Comment | May 2, 2009
  • I’m also confused about your hype cycle graph. When i looked at it I expected the “augmented vision” to be in the plateu of productivity. I would agree with your written portion declaring that we will achieve overhyping with markerless AR. It is afterall, very gimicky and I haven’t seen that many real useful applications.

    I don’t think RFID has a place on the cycle. I recently posted about it - http://thomaskcarpenter.com/2009/04/15/300-billion-rfid-tags300-billion-rfid-tags/

    Realistically I think object recognition will move past any need for RFID tags, so its already obsolete.

    For me, I think an AR development cycle really has to be broken down into the subcomponents (similar to what Rouli has said). The rate of improvement on each of them will really dictate the pace of implementation within greater society.

    Interesting discussion, keep it up. :)

    Comment | May 3, 2009
  • @rouli

    Interesting two dimensions. I am just thinking how we can put these different examples of hyperlinks and interfaces on a sliding scale.

    Hyperlink: dumb —> smart? (mapping dimension)
    Interface: isolated(parallel) —> integrated(mixed)(augmentation dimension)

    Comment | May 3, 2009
  • This thread is almost getting better then the post. I like. Forgive me if you think my replies are slow. I take my time to think about them and to keep the quality up.

    @rouli Good that you understand the hype cycle now. And thanks for the mention on Ori’s and yr own blog.

    @Thomas K Carpenter The x axle is used to portray the time between the technology trigger (the invention) and the general application in every day use (in any form, also industrial, as long as it generate value). Therefore the ‘newest’ thing is on the left, not on the right.

    Please see http://en.wikipedia.org/wiki/Hype_cycle for an explanation on how to interpret the hype cycle

    @rouli As for the two dimensions. Good idea.

    I would generalize the 2 axles: hyper-linking technologies is to narrow (GPS wont fit there) and interface maybe too. I do like manner of linking vs manner of augmentation by @raimo.

    Lets see if I can explain and extrapolate on this:
    A dumb link is a barcode which the ‘computer’ translates into a link. WIth this the computer still doesn’t know anything and you need to click and open a browser to do any thing intelligent with it.

    In other words: the link is dumb. This is also an example of a simple interface (as @raimo calls it) or simple ‘augmentation’. Actually the augmentation is not really there. Visually nothing is rendered.

    The same goes for my face being recognized (as in iphoto) and my name is displayed. This is a dumb way of recognition and use of ‘augmentation’. The device only knows my name nothing more.

    A smarter link is used with the WWF China bear. The device recognizes some patterns (lines, movement?). Reality is the marker here yet the computer just looks for simple patterns. No location is used for instance. The subsequent augmentation is complex yet seems 2D. The level of ’smartness’ of the link is determined on the amount of sensors being used (marker, location, movement, compass etc).

    A smart link is when my face is recognized (who i am) including where plus my status, who my friends are in that context (location plus any social network). This is then displayed on the screen. You will see me, my status as text and my ‘friends avatars’ in the neighborhood, sized and related based on distance to me. So the device knows more then just my name and displays it in a rich way.

    On these two axes we can put all current and new technologies in a smart meaningful way. Lets call it the Augmented Definition Matrix (I like names…).

    Comment | May 4, 2009
  • Here is the first draft of the Augmented Reality Definition Matrix http://www.flickr.com/photos/maarten/3500834851/

    Comment | May 4, 2009
  • I like it! It really puts things to order!
    I like it so much, I think I’ll steal it and use it on my blog :)

    Comment | May 4, 2009
  • Excellent graph. As an engineer I’m always impressed with content filled graphs. :)

    Going to think about it some more and then I’ll have some feedback on my blog as well. I’m thinking your definition of smart and dumb links might be better as:

    Dumb link - information portrayed at face value (nothing added) without considering the enviroment the object is in

    Smart Link - information used to create new and interesting relations to other information(multiplying the information) that considers the enviroment the object is in

    Good discussion. More to think on.

    Comment | May 5, 2009
  • Nice work. I enjoyed the analysis and frankly “the hype” about AR. I would like to add though that it is quite possible that the Augmented Vision piece may be much closer than 3 years out.

    Vuzix not withstanding there is a lot happening on this front in the world.

    Is there a way to be stay in contact with everyone interested in AR as the future unfolds? Vuzix clearly has an interest in promoting the topic and we have some great augmented vision products coming for those with interest.

    I hope that this input does not come across as “promotional”. It was not meant to be.

    Comment | May 18, 2009
  • Laurent

    Hello,

    Thanks for this great analysis. What you started here was needed. The search engine miracle helps to find you, i don’t know how exactly.

    I just started this week to focus all my personal and business time on the AR market the ‘Smart Link’ way. I bet my start-up business on this Augmented Vision.

    Is there a way to be stay in contact with you all and share views ?

    Paul Travers, I take the opportunity to ask for a chat, no commercial way, but more as business vision shares.

    Regards,
    Laurent

    Comment | May 20, 2009
  • [...] of the user’s location and other context to superimpose relevant information. In the Augmented Reality Hype Cycle, Maarten of SPRXMobile positions the four styles and other additions on a hype cycle chart, [...]

    Pingback | May 29, 2009
  • [...] Interesting posts from a variety of blogs: AR Hype Cycle, Metaio’s AR Software Demo and High Visual Quality AR Demonstration.  I’m certain [...]

    Pingback | June 5, 2009
  • [...] Full article: The Augmented Reality Hype Cycle [...]

    Pingback | June 12, 2009
  • Great post, and nice segmentation analysis.

    Personally, with iPhone 3GS I think that there is a really interesting play to leverage the new hardware accessories support within iPhone OS 3.0 to build external 3D/VR glasses which plug into the iPhone, something that I blogged about in:

    3D Glasses: Virtual Reality, Meet the iPhone
    http://bit.ly/141NNg

    Check it out, if interested.

    Cheers,

    Mark

    Comment | June 17, 2009
  • [...] SPRXMobile Mobile Service Architects » The Augmented Reality Hype Cycle via kwout [...]

    Pingback | June 21, 2009
  • [...] I haven’t seen a great deal of what might be described as that. Others have, though, like Maarten Lens-Fitzgerald writing about the AR hype cycle in April: This year Augmented Reality (AR) is out of the box. It has escaped from the universities [...]

    Pingback | July 27, 2009
  • [...] SPRXMobile Mobile Service Architects » The Augmented Reality Hype Cycle via kwout [...]

    Pingback | September 14, 2009
  • [...] as it’s likely to be an evolution of the mobile social web. While a few years out, see the proposed Hype Cycle, let’s spend time thinking about what the future could [...]

    Pingback | September 20, 2009
  • [...] learning from Robert Rice and Dave Elchoness to see  how it develops. While a few years out, see the proposed Hype Cycle, let’s spend time thinking about what the future could [...]

    Pingback | September 21, 2009
  • [...] learning from Robert Rice and Dave Elchoness to see  how it develops. While a few years out, see the proposed Hype Cycle, let’s spend time thinking about what the future could [...]

    Pingback | September 21, 2009
  • [...] learning from Robert Rice and Dave Elchoness to see  how it develops. While a few years out, see the proposed Hype Cycle, let’s spend time thinking about what the future could [...]

    Pingback | September 23, 2009
  • [...] learning from Robert Rice and Dave Elchoness to see  how it develops. While a few years out, see the proposed Hype Cycle, let’s spend time thinking about what the future could [...]

    Pingback | September 23, 2009
  • [...] from Robert Rice and Dave Elchoness to see  how it develops. While a few years out, see the proposed Hype Cycle, let’s spend time thinking about what the future could [...]

    Pingback | October 2, 2009
  • [...] apps) and may even create a backlash on the AR concept, making an earlier than expected dive on the hype cycle. I already went through this in the late 90s when the hype was all about virtual [...]

    Pingback | October 2, 2009
  • [...] apps) and may even create a backlash on the AR concept, making an earlier than expected dive on the hype cycle. I already went through this in the late 90s when the hype was all about virtual [...]

    Pingback | October 2, 2009
  • grendelan

    Some things are so obvious they don’t need a name. Some things are so silly they shouldn’t have one. This seems to be both.

    Comment | December 21, 2009

Leave a comment

Powered by WordPress. Theme: CKM Extended based on TheBuckmaker