I am trying to figure out how nokia point and find works?
For instance, when I scan something does it send the information into a server for resolution?
And then once resolved, send back the content?
Hi Kevinnokia,
Recognition in P&F happens in real-time, on the client. There is no call to a server (unlike most of our competitors... he-he), so it's much faster to get the information you need.
The directory is also on the device, so here again, the search will happen in real-time.
Regards,
SC
P&F team
It seems like that would limit point and find greatly...it would seem like no more than a glorified bookmark!
The content or url or pointer is already is on your handset somehow??????
also, how can you have that big of a directory on your handset? ????????????/
RE: scmoatti
"Hi Kevinnokia,
Recognition in P&F happens in real-time, on the client. There is no call to a server (unlike most of our competitors... he-he), so it's much faster to get the information you need.
The directory is also on the device, so here again, the search will happen in real-time.
Regards,
SC
P&F team"
What you are saying confuses me. In essence you are stating that all media references for anything a user could "point and find" at is locally stored on the users phone? That would require an HUGE amount of local memory on a personal cell phone. Last I check they have yet to make a 200+GB sd card let a lone a cellphone that can utilize one larger then 32gb. Either you are not fully explaining the process or you yourself are miss-informed. For technology such as this to work correctly an image would be taken via the client side, that image would then be referenced for likeness to other images (for example a picture of a movie poster could be referenced to a google image or something off IMDB).
I will use an example of what I think your real-time is. I will use the application POYNT (for blackberry's). Although this application does not link the real world to the virtual it will work just the same.
I can search using POYNT for say a movie. That search will populate using various sites on the internet various results in relation to my current location via GPS. Once the results have populated, I can view theaters and show times in my area, movie reviews, a theatrical trailer and much more information related to said movie.
Point and find MUST use an indirect look up process for it to function and for (from what I am reading) you to claim otherwise is completely wrong. Even further, what would be the point of Point and Find if Nokia could not track what is being searched? What would be the point of all of this then?
This is how I believe point and find TRULY functions.
P&F -> central database -> specific location(s) -> results -> central database -> phone/device
1. The central database would have to have some sort of function to put different searches into various categories to point them in a direction
2. Those categories would have certain locations that they would look at for results
3. Those results would either a) be from a mobile friendly site or b) would be redirect back to the database to be re-indexed into mobile content. Regardless of A or B once a search has been redirected back to the central database it would NOW be locally stored on your database so it would cut the time from search to result by about 75%
4. Finally the results are populated on the users phone/device
Please feel free to respond, and don't be scared to use technical terms. I have 15 years in coding and 10ish years coding for windows mobile and java based OS's for mobile devices so I am sure I will understand you.
Thanks for your time!
After doing some reading, this is what I came up with.
Actually this is EXACTLY how it works.
1. A client creates an account with P&F and associates their device with the account.
2. Client takes pictures of the content they wish to advert and links them to their account.
3. Client assigns 'link' to those objects via the management portal on the P&F website.
4. Consumer takes picture of something.
5. That picture is referenced with the P&F database that contains all the clients.
6. If there is a match, results are returned.
First, the client themselves link real world objects via their device to a account profile they have created on the P&F database. Second the user links images from their device to a clients profile on the P&f database.
See here
http://pointandfind.nokia.com/docs/PAF_Management_Portal_intro.pdf
Page 4
1. Get started
Create and name your Nokia Point &
Find world using the Management
Portal
Hi ziptied,
Well, you really did a lot of research on P&F and we here at P&F feel quite honored that you have taken the time to digg into the technology so much... Thank you!
You're perfectly correct that anyone can create a custom experience via our management portal. In addition to the brochure you referred to, we are preparing a video tutorial that will explain how to create such experiences easily. It will be posted in the next few days at this link: http://pointandfind.nokia.com/?login.
On device, once you have selected a use case in the carousel, the recognition happens in real time, without the need to take a picture or make a call to a server. If you are in the US or UK and have a chance to try P&F on a device (S60 3.1 only for now), you will see why it can only be so: server-based recognition typically takes anywhere between 10-30 seconds, depending on the quality of the network connection. With P&F, you get results in just a few seconds.
I hope this helps answer some of your questions and again, we in the team really appreciate your interest,
Regards,
SC
P&F team
What about this part
"This is correct? If so, then when you stated competitors above, who exactly where you referencing? I am assuming it is not Neomedia technologies and that you are actually working with them due to the fact they do hold the patent in both the US and in europe for indirect applications from the real world to the virtual. Is this safe to assume?"
Regardless, the process still used the indirect look-up method.
Thanks!
In addition, when you are talking about the carousel, I assume that once the category is selected then the server pushes references? Like I said earlier, how could it recognize anything only on the device itself. The memory that would be required is immense.
I understand the concept of recognition technology. Common points are found and referenced to a likeness. For instance facial recognition, common points are referenced and then indexed and bounced off a repository. How is P&F different?
Hey ziptied,
I don't have enough understanding about the subject matter to reply to your question, but I just wanted to let you know that most discussions about business/legal/IPR/partnering/etc issues are out of scope of this discussion forum - especially for "Nokia Staff". Hope you are cool with this!!
cheers,
- tommi
I am cool with it! For the director of Beta Labs to personally respond to little ol' me with the way you did 6 days after my last post puts a smile on my face. I must of sparked a flame.
Cheers!