Here at Dahu Towers, we are always on the lookout for the next big thing, or even the next little thing if it manages to amuse us for long enough. And our attention was recently drawn to an intriguing new app for augmented virtual reality, namely Aurasma, the latest plaything from Autonomy, themselves the latest plaything of HP.
We first saw the Aurasma visual browser at the recent London AdTech show where it was introduced by Matt Mills from Autonomy. In an impressive live demonstration, he photographed a page from a glossy magazine, which he then uploaded to a central server along with a short video taken of the audience. Within seconds, he was able to point an iPad running the Aurasma app at the original magazine photo and have it recognise the image. Presumably the app was sending a series of live input images to a central service which searched over its library of tagged images in real-time. Once it matched the newly uploaded photo, it immediately linked, downloaded and ran the video clip showing the awed AdTech audience in place of the original magazine advert.
Very impressive on several levels. Firstly, just to run a live demo at a large show over a relatively slow 3G link, that involves uploading two multimedia files to a remote server, shows an impressive level of confidence in the entire technology stack, from iPhone through dongle to Aurasma’s back-end servers and back to the iPad running the app. Big respect to Matt for having the cojones to attempt that!
Also one has to be impressed at the speed with which the uploaded image was received, processed, indexed and made available for searching by the back-end services supporting the application. Presumably, the search is based on Autonomy’s IDOL platform, which, in our experience, has not always proven to be the most scalable or performant search engine in the market place. One does wonder how it will scale as the number of adopters of the free downloadable app increases. For now though, it works impressively well.
We were left wondering as to the uses for this technology. Autonomy’s Mike Lynch claims this is the idea that will make augmented reality go mainstream. As a consumer app, it does seem to offer huge potential for fun and amusement to comment, tag and share your view of real-world objects and places with your friends (although the latest comments on the itunes app store from early users would suggest some further work is still required on usability). From a commercial perspective, which we think is where Autonomy expects this technology to go, we are less convinced. The idea is that static, 2-D advertising can come to life; point your smart phone at a newspaper and see the latest news headlines in place of the outdated printed copy; link the label of a jar of pasta sauce to a recipe; take a photo of a public building like a major gallery and see what events are happening inside right now, with links to buy tickets.
One can see committed geo-cachers might appreciate the new world of possibilities they now have to hunt for virtual satisfaction. For the rest of us, do we really want to see the world through the lens of a smartphone?
More worryingly, we don’t see how this would work in practice. At the AdTech show, Mr Mills appeared to select a page at random from a glossy magazine. From memory, the page showed a large and presumably expensively prepared advert for a perfume. I don’t recall any permission being sought from the owner of the image rights before Autonomy tagged their prime advertising space with an amateur video of the AdTech audience. I’ve had a quick look around the Aurasma web site and I do not see any options for disgruntled victims of Aurasma graffiti to object and have images removed. I’m sure the lawyers at Autonomy will have considered this scenario before launching the product, and clearly there must be protection available to the holders of image rights and copyright, but from the information available on the web site, I do not see how it is intended to be controlled.
Next, one imagines the most popular images may well be tagged as “auras” multiple times, from different users or companies. Is there to be a market for the top position in an Aurasma image search? If so, how is this to work, and how is it to be governed? Are we opening up a whole new world of SEO for images? A new market for auctioning images and faces on which to stamp messages and content by the highest bidder?
Imagine a competitor taking a photograph of the entrance to a corporate HQ, for example Autonomy’s in Cambridge. They can upload their own messaging to appear in connection with that building. Imagine an anti-capitalist organisation tagging the Palace of Westminster with anti-globalisation messages, presumably to remain associated with the seat of Parliament for all time? We understand “auras” will soon be possible based on a face. Imagine if your face were to be tagged with a link to offensive or even pornographic material. How is one to demand instant retraction of offending images? Is there to be a hotline to Autonomy to report and remove offensive, libellous or illegal tags? As I say, I’m sure the Autonomy lawyers have thought this through, but their preferred remedy is not clearly identifiable on the Aurasma site.
Interesting, but for now, we just don’t get it.