GraffitiGeo shows first augmented reality app for restaurant recommendations | Hacker News Reader

GraffitiGeo shows first augmented reality app for restaurant recommendations

graffitigeo.posterous.com

79 pointsjmtame37 comments

Oh no... now along with the bluetooth people talking to themselves, you'll have a bunch of people scanning slowly around with their iphones out.

But seriously, this is awesome. Great work!

I have mixed feelings about 'augmented reality' (especially with regards to those bluetooth jerks), but this seems like it would be extremely useful with HUDs.

Maybe AR is new to you, but AR isn't new to me (I worked with video overlay AR in college in 2002), and I fail to see why anyone should care about these mobile phone AR systems. Cell phones have poor video quality and frame rate leading to a mediocre video overlay system at best. Combine that with the screen space wasted on a street level perspective, and a shaky hand, and you have an application that sucks all around. Regardless, I have to watch people pat these developers on the back as if it were difficult to combine a GPS signal with accelerometer and compass readings and project some boxes into a 3D scene.

Please take this with a grain of salt: nobody cares. Users want an easy to read, easy to navigate map, with good review data. The built in Maps application nails the first two (it even uses the compass), and does an okay job on the last one. You should be concentrating on what isn't already done well.

I agree, but didn't want to be a complete downer since these guys no doubt worked hard on the app and are emotionally invested. (I'm also a hyper-critical person, so I try to censor myself. My wife would have divorced me already otherwise.)

Maybe this first iteration is really just a starting point for them, and they will rethink their approach? Releasing something is more than most people do. I think some of the praise they're receiving is because people respect their effort and their GENERAL conceptual direction.

I certainly don't want to judge based on a first cut, but I don't see any of the fundamental problems solved. AR has been around a long time, and has never been as everyday popular as first impressions have suggested. Show me usability improvements, a compelling subject matter, or at least a superior implementation given the hardware limitations. I see none of that here, just another echo of flashy AR demos (http://news.ycombinator.com/item?id=670645). To me, this is the result of web programmers hearing about augmented reality for the first time, or realizing that the iPhone is the first everyday computing device capable of AR, and hacking something together over the course of a couple months.

Agreed.

Like I said, I'm just in a 12-Step "Debbie Downer" program, so I'm trying to be more positive. Baby steps ;-)

It's cool that people are pulling this off, but I can't help but wonder if they're asking the wrong question: "can we?" instead of "should we?"

Maybe this is just meant as a proof of concept. In that case, it's great. But if it's meant to be a useful way to find restaurants and get reviews, I think there are much better interfaces for that. Including existing apps.

It is not about restaurants. Imagine this scenario:

When you are walking on a street, you pull out our app and look around you. Now, you know everything about that street.

It is a totally different experience if you are trying to find this kind of information on a 2d map. It is like google street view vs google map.

Imho 2d google map is superior interface for finding information. I don't need to move with my phone around to get information etc. Was it your need to be able to recognize buildings based on street-view, because you can not read it from map? Technically impressive, thumbs up.

Absolutely. AR is probably the hottest thing now in the valley riding on CrashCorp, etc. So will AR be useful as a new way to find reviews and restaurants, who knows. Time will time.

It may be a good way to find VC money and like minded people, but it's totally useless to the rest of the world.

I think you are correct mrshoe (regarding the "can we" vs. "should we" and also about using the best interface for a given purpose.)

We developed an app very similar to this (restaurant ratings, school ratings, vacant apartments, homes for sale, home sale prices, traffic flow visualizations, etc. etc.) for our real estate brokerage and real estate investment division. It got a nice "gee whiz" response from some of our geekier clients, but that was about it. We stopped using it on phones when our first client nearly broke their leg by tripping on a curb. [We don't like lawsuits.] We now use a multitouch tablet instead. The picture quality is much improved, clients can use gestures to manipulate the data, and they don't run into things anymore.We just call them 'interactive maps' so we don't scare our less tech-savvy clients.

This type of thing has been around for awhile. I don't completely understand the novelty this holds for this particular (techy) crowd. I guess the future is still just really unevenly distributed...?

I love seeing these AR apps, but being able to integrate with some kind of lightweight head mounted display will make them vastly more interesting. Anybody know what the state of the art on those things are yet?

if anyone out there is building this, we would love to extend our api recommendations to you -- contact me if interested (jared@graffitigeo.com)

AR has been done for ages with head mounted displays (lightweight is mostly a factor of time+money the AR rigs or 80's were not light but the ones from 2000's are)

Google "Wearable Computing", "Professor Steve Mann", "MIT's Media Lab". http://www.eyetap.org/

Can someone explain how this works? Do they use the GPS + compass to detect what you're looking at or do they actually analyze the image data. I'm thinking it's the former.

Probably GPS + Compass + Accelerometer. You need to be able to detect the angle of the phone as well (looking up, level, or down).

The iPhone doesn't have the processing power to do meaningful real-time image recognition.

I would agree with you - but the 'infobox' (for lack of a better term) stays stationary over the same point of the building for the whole thing. I can't imagine the computations to keep it that centered based simply on the orientation of the phone would be either accurate enough or much faster than basic block/rectangle image processing.

After all, Google street-view does some simple image processing like that in-browser using JavaScript, and I'd imagine native iPhone code performance would approach that level of power.

With the accelerometer you can detect gravity (true down) and your angle relative to it. Then model the earth as a plane and keep the label a constant distance above the ground.

Seems like they did a good job of it though. It's really slick.

If they were doing feature tracking (following recognizable points in the video), this would far smoother. They don't even appear to be filtering, as their info boxes are shaking all over. Anyone demoing an AR app who moves the camera slowly is compensating for poor implementation.

GPS for location, compass for direction. And it'll only work reliably outdoors since there's no direct access to the camera stream.

Awesome. I hope it works fast though I wont mind hanging on it for couple of minutes if the results are good.

How is the location accuracy - one can potentially focus on some thing a 100 yds away. So would it still identify the correct restaurant or will it list n options.

ok, Is vidwo upside down just for me? If not, wtf?

Upside down for me too. Seriously weird.

This is with Movie Player on the latest Ubuntu.

Video looks good to me.

i also uploaded a version here: http://www.youtube.com/watch?v=BeuL94w1G8E

seems fine for me too. Guess Linux isn't ready for the desktop yet!

All these AR apps get a lot of buzz/press but in the end they use unsupported APIs and won't get approved. Worth the PR perhaps.

Awesome, especially if it has alot of reliable reviews. It would be great to just stand on the street and look around you for a good place to eat instead of browsing through yelp or zagat reviews. Great job guys!

we currently collect data from both our own iphone users going around and voting up on places, and from large and credible third party recommendation sites. you can get a heatmap representation here and see if it passes your own test: http://graffitigeo.com/heat

Not exactly the first - Zagat released one a couple months ago: http://www.zagat.com/Discuss/ForumPosts.aspx?SNP=NFZE&TI...

Albeit a slightly different variety of AR (using the compass and GPS, but no camera. Still fun). Perhaps there are others?

awesome. have you tried it in san fran yet? i walk .5 a mile to the bart all the time and it would be cool to see if there are some good restaurants i'm missing.

reminds me of this http://news.cnet.com/8301-17938_105-10266380-1.html

Cool demo but again. Will people adopt this technology? Who knows.

What happened to the Sekai Camera from Tonchidot? Just vaporware?

i love this so much esp in busy cities. very cool guys!

so freaking cool. congrats guys.

Reply on news.ycombinator.com