Posts Tagged google
Eye-Tracking Will Be The New Click-Throughs
Posted by Tom Carpenter in augmented reality on March 22, 2010
Part of the Internet economy is built upon the “click-through” or CTR (Click Through Rate.) The CTR attempts to measure customer interest in a particular product. If a person finds the banner ad interesting enough, they will select it and be sent to that site, hopefully to purchase a product (or Conversion Rate.) Thus the effect of the advertising can be measured and billed.
Even the layout of a site can affect the conversion rate. Or in the case of Bing, Microsoft’s new search engine, the color blue can be worth $80 million in additional usage. Website optimization rearranges the layout to achieve fung shui for dollars. What is measured can be improved.
These products, like Google’s search engine, are worth more than just the product itself. And they get first crack at the wealth of information flowing through their servers. Using the misspelled words on Google search, they created the most robust spell-checker on the planet, in every conceivable language.
Using the information of what you click, they can run experiments to see what works best. Collaborative filtering makes recommendations to users based on what other users like. The data exhaust of websites can be as valuable as the product itself.
As augmented reality products use eye-tracking to achieve a realistic virtual overlay like in the recent GM augmented windshield, they are getting more information than just how to align the graphics. Eye-tracking adds a new dimension to the data exhaust. As any professional poker player will tell you, the eyes are the window to the soul, and to the tell. Someone holding pocket kings might look down at their chips the moment they see their cards in anticipation of seeing a bigger pile later. Players wear glasses for a reason. The eyes can give away important information.
Studies on select groups of people using eye-tracking have given broad generalizations (read the before-mentioned link for more details):
1.Headlines draw eyes before pictures.
2. People scan the first couple words of a headline.
3. People scan the left side of a list of headlines.
4. Your headline must grab attention in less than 1 second.
5. Smaller type promotes closer reading.
6. Navigation at the top of the page works best.
7. Short paragraphs encourage reading.
8. Introductory paragraphs enjoy high readership.
9. Ad placement in the top and left positions works best.
10. People notice ads placed close to popular content.
11. People read text ads more than graphic ads.
12. Multimedia works better than text for unfamiliar or conceptual information.
Imagine what can be learned when the eye-tracking is always on and always sending data back to the home servers. Contextual filtering will “pigeon-hole” you into a type of viewer and give you a website more suited to your style of reading. Web design will customize based on your changing eye sight. Older viewers that linger over the words will get larger fonts so reading isn’t so strained. Colorful pictures will attract younger viewers so advertisements will be changed to align with them.
And that GM Augmented Windshield? If their sensors can identify advertising along the side of the road in the form of signs and billboards, then can they collect the data on what works and what doesn’t and sell that to ad agencies.
Like I said, the data exhaust can be more valuable than the data itself. And eye-tracking will prove to be more valuable because its an unconscious reaction. Just be careful where you’re looking.
Popularity: 100% [?]
8 Things That Are Possible With Google Goggles
Posted by Tom Carpenter in augmented reality on December 8, 2009
Unless you’ve been sleeping under a pile of old Commodore 64s, you’ve heard about the Google Goggle announcement. The service allows you to use pictures to search the web. The service is amazing in its own right. Check out the videos if you haven’t seen them already.
Google Goggles opens up huge possibilities for augmented reality. One of the challenges of AR is for the computer to understand the world around it in a realistic manner. GPS and compasses only tell one tiny part of the story. If the only thing you know about the world is the exact location you’re in and which direction you’re facing, then you really can’t accomplish much.
Object recognition (which for Goggles includes facial recognition, but they’re not allowing individuals to be accessed until privacy issues are figured out) allows our phones to mimics how the brain works in the seeing the world around it (or the Reality Recognition if you’re familiar with my AR Scale.) This doesn’t mean that Google has mind-mapped our mental processes to make the service work. What it means is that Goggles allow for computers to start doing what our brain is doing–seeing the world around it.
Google Goggles Will Make These Eight Things (and More) Possible:
1 – Cataloging your refrigerator and pantry so you can cross-reference at the store and get menu ideas. Maybe in five years your fridge will have a camera inside of it to do it real-time.
2 – Noah Zerkin on twitter pointed out pretty quickly that “If they can recognize scene elements, and they also provide GPS/inertial/mag-based overlay, how long before using surroundings to fix POV?” This will make immersive augmented reality games and services possible.
3 – Run your TiVo through it so it can edit out the commercials.
4 – Wilderness survival tool that lets you know if its safe to eat the purple berries (when the plant picture search is added and assuming you have cell phone coverage!)
5 – Scan your strange garage sale items to find out what they go for on Ebay (Who actually owns purple zebra lamps?)
6 – Warehouses or stores can auto-count items as they are removed from the shelves with a bank of cameras so they can have real-time inventory and also know which items were “almost” sold.
7 – Genealogists can use old family pictures as a search criteria to find other long lost family branches.
8 – Trend spotters could use live camera feeds to understand usage patterns of products.
These do require access to Goggles database or at least a way to port through Goggles (like using Twitter for real-time search) and some might also require usable HMDs to make it worth using, but the possibilities are exponential. While none of these things are going to happen overnight, Google Goggles sure has made things a lot more interesting (especially if they allow facial recognition.)
Popularity: 16% [?]
ACME – One Piece of an Augmented World
Posted by Tom Carpenter in augmented reality, digital singularity on October 12, 2009
The augmented world will exist as a shadow to the real one.
Rouli from the Augmented Times posted on Friday about the ACME project (Augmented Collaboration in a Mixed Environment.) Leave it to a group of researchers to suck the life from a wonderful tool by giving it such a drab name as ACME. And I disagree with Rouli’s assessment that it’s a form of telepresence, its much cooler than that. My point is not that their descriptions are lacking (and I have nothing better to offer). It’s more that the technology deserves sexier nom de plume.
The video demonstrates the idea that presence need not be tied to our geospatial location. Nor does our awareness. An augmented world will allow us to move the focal point of our being to a different location.
Though the word demonstrate has limitations. The Mixed Reality Teleconferencing (see english website for more details) ACME project shows us how we can mix reality and the virtual in a boring conference room using Second Life. Their skill at manipulating these two environments is commendable but unfortunately, Second Life is a not platform that I believe will be a part of the equation in a real open AR world.
The reason I think Second Life doesn’t work is because it is a seperate world, similar, but unlike ours. It has virtual ground, trees, buildings, and people, but not in the same location as our own. Google Earth gives us a 3D representation of our world in the exact proportions we need it. A GPS position on Google Earth is the same on the real Earth. Second Life doesn’t have that symmetry with our world.
Another connection between the augmented and real world can be made using the technology demonstrated by the researchers at the Georgia Institute of Technology as they use cameras to integrate real world activity onto Google Earth. They will also be attending ISMAR09 in a few weeks (maybe these two groups should collaborate?)
The combination of these technologies and a lightweight HMD can give us a way to project ourselves to another location and be aware of changes in that environment. This will create a world in which time and distance have less meaning. And while its not going to replace the feeling of walking the streets of Shanghai or exploring the sand dunes around the Great Pyramids, it will certainly make the world feel smaller.
This augmented world will connect people and places in ways never considered. Space will be layered with the human spectrum, games will exist across imaginary dimensions and the reach of awareness will be pushed out to a global scale.
Some may scoff at such thoughts and believe that an augmented world will only create a populous lost in an unreal world suffering within its own delusions. I offer the doubters this video as proof that “reality” is just a trick of our monkey brains.
Popularity: 12% [?]
10 Things Your AR App Must Have to Succeed
Posted by Tom Carpenter in AR Games, Steal This AR Idea!, augmented reality on August 23, 2009
With the iPhone OS 3.1 launch set to debut next month, I wanted to explore the features that are going to help make some AR apps succeed over others. We’re going to be seeing a spike of releases since many have been working towards this moment waiting for the iPhone to become AR enabled, so its a good time to get those apps in tip-top shape.
And while I’m personally smartphone agnostic, these are features that should be present on any app, no matter what the platform. But the OS3.1 release will connect the hype of AR with the hype of the iPhone, so that’s why we need to talk about app features now. Call it Hype2.0 if you’d like.
And now for the features:
1 – Solutions Not Gimmicks
The best products and services are solutions to a problem. Going to your webcam to open up a 3D version of the product has lost its luster as a tired gimmick. Find a reason that augmented reality can be useful for your customers. And if you’re a car maker looking to use AR, listen to Rouli.
Don’t be like Always, the feminine hygiene company, please…
2 – Social Connections
The reason these social connectors like Twitter and Facebook work is because they allow us to be nosy neighbors seeing what next person is doing. The term “Keeping up with the Jonses” applies here. Give your app a way to connect other people in creative and interesting ways. Even if its seeing the high score along with a picture of you dodging a cannonball.
3 – App Search
Google dominates the search wars because they’re the best at it. Many functions of AR require pulling information from the data sphere and if it gets hijacked by spammers then the app will quickly lose its purpose. No one wants to pull up a restaurant layer to find a nice place to eat in a new town and get a porn layer instead while your three smiling children peer over your shoulder.
4 – Alacrity Wins
I could say apps need to be quick, but they need to be more than that. The Webster’s definition of alacrity is “promptness in response” or “cheerful readiness.” Apps should anticipate what the user wants and get it quickly. If it takes 10 seconds for the camera to identify where its at so you can decide which way to walk, no one is going to use it. If the processor is too slow, move your heavy lifting to servers off the smartphone. Life moves quickly, your app should move at the same pace.
5 – Unique Style
If you want your business to grow, your style should be tasteful and easily recognizable. When people see your products they should instantly know who created it. Already the Layar symbol has stuck into my head as a good product identifier.

6 – Not Another Vista!
If I let my kids cook dinner for themselves they make cereal. My wife will take six hours to make an exquisite dinner that uses every dish in the house, including a few of our neighbors. The kids eat cereal because its easy to do. My kids are your users. They’re not going to make Chateaubriand no matter how good it tastes. So make your apps simple to use.
7 - Save the Trees, Please
Please don’t make me print out another marker to see your version of augmented reality. We might be able to say that AR is a green product, but for all the papervision markers we’ve had to print out.
8 – Now With Less Spam
For the AR apps that allow user added content, we need to make sure we’re not besieged with spammers like the regular Internet. While this might be an impossible request, at least consider the spammers in your product design.

9 – Tell Me A Story
Augmented reality was built by programmers, but it needs artists to grow. Don’t forget to tell me a story and engage my emotions. AR is a wonderful new medium with potential to tell stories in ways never told before. The story can exist all around us or even in our living room.
10 – FUN!!!
Now matter what the purpose of your app, even if its a serious one like saving the rain forest. Don’t forget to make it fun (and maybe include a little magic).
Popularity: 62% [?]
Leverage The Crowd To Fill The Cloud
Posted by Tom Carpenter in augmented reality on June 23, 2009
Augmented reality is only a medium (actually just a mashing together of other media), but content is king.
To make use of computer data and images superimposed on real life we need content, but creating that content can be time consuming and expensive. Instead, we have to find ways of leveraging the crowd to fill the cloud. As I’ve explained before, cataloging image databases will be one way we can build digital versions of the world.
Today, researchers at Google are presenting a paper on Computer Vision and Pattern Recognition (CVPR) in Miami, Florida. Using 40 million GPS tagged photos from Picasa and Panoramio and online tour webpages, they’ve been able to improve computer vision of major landmarks. The technology sounds similar to Microsoft’s Photosynth, but they may have different implications based on the form of the data.
The key point is that these technologies (either Google’s or Microsoft’s) will allow for the leveraging of information (in the form of pictures) being created daily and stored on the web. Not only is this data available, but due to its time signature, it can also help us reconstruct past events of significant importance. This will give us powerful tools for creating huge chunks of content for the cloud.
The below picture shows the visual representation of how their cluster recognition model works.

Google has shown us another way to use large, noisy datasets to automate the digitalization of the world.
Via Spatial Sustain.
Popularity: 1% [?]






Recent Comments