Hello people,
I was brainstorming on a 'new' idea with a friend on a boring night we referred to as a 'borinthon' _{(mimicking Facebook's hackathon)} in my room when i came up with a crazy idea to create something like an application/ website that looks like the face detection technology but in a different format. The face detection technology i know about is used in some websites like Facebook to match photos of people in that website but what of that which sources for photos all over the web and which may not necessarily be human face, does it exist
It's going to be like this for example: I would upload a picture of someone or anything like a 'thing' and then search for it and the result would be pictures of people or 'things' whose face matches the one i uploaded. I could either upload part/ half or quarter of the photo to find which one matches it for a full size. The result would be gotten from different web pages and content which hosts images or image-related contents available over the internet. This is similar to typing something into Google search lemme say Ajibel and the return page would contain results from web pages which has the keyword Ajibel in it. _{Type 'Ajibel' into 'google' now and ego-surf me!}
Therefore this idea is for matching of photos over the net.
You can use this method
-to know the websites or the profile of a particular photo you uploaded to match/ search for (a form of stalking when used negatively)
-to get matches of a quarter or part of a whole photo you've uploaded
-get related/ suggested photos which matches the photo you've uploaded
-to make comparisms and similarities with the photo you've uploaded and your searched results
-to make bla bla la la etc...
-and some other stuffs i might not have thought of now

I nicknamed this 'Fotogle' (process for searching- fotogling , people who search- fotoglers ) and this site would strictly be for uploading of photo search matching. That may not be the name afterall

This may also be used to find someone you may have lost touch of since childhood by just uploading the picture of the person you have while a kid and you could get the match of the face of such person who may now be an adult or a granny or dead

Why do i Post this? My fear was that such could be existing already and so i need to think of another idea but i felt uneasy as part of my mind suggested i should voice it to know what others think of it before i clarify my stance that such platform like this exists
When i described this to my friend, he thought me to be mad _{(madness is a gift)} but he later got a bit of sense in it. The problem is, we both dont know how to start off and another thing is, we dont know if this our 'new' idea is already existing because 'No Idea is a new Idea'.
So i posted this to know if we already have such an application or a website strictly dedicated to it or if not, is such an idea possible to get off the ground- what resources or skills are being needed? Can this be possible? Is this crazy idea caused by a boring long night with no sleep? Many questions could follow if i dont stop!
I know of PHP, i've been learning it on w3schools but i cannot boast to be a programmer in the sight of geeks who programme with eyes close ... and i know to create such, one needs to have command of many programming languages. I've 'googled' and 'goooaled' on this idea but i'm not finding the best answers even on 'yahoo answers'. I've also posted this on some other forums so i can know the minds of many but some are just reading with no comment preferring to view my profile and see the picture of the silly crazy poster or saying goodluck good luck. If this is a new idea, then i must have gifted a 'bored or jobless' programmer with a beautiful idea (i'd be happy and sad if someone would steal the idea, i may not mind) if otherwise.... i'd also be happy this idea doesnt get off the board so i can concentrate on my website
But even if something like this exists, someone amongst every nairaland programmer could create a different iteration and work with another someone to think of more 'new' ideas to support it. The whites have always accused the blacks of copying ... i didnt copy this idea , i stole it using the face detection technology methods
'Lesser minds copy, Great minds steal'
Anyway, i'd appreciate your replies in form of suggestions , commendations , criticisims and condemnations

jus that every image has a name and people wud readily google that name than the image. Unless they are sure searching with the image will yield better results than the name and this means that your image detection algorithm will have a 95% success rate and your search engine can search millions of databases using that algorithm very very very fast...

Your idea is not really new, it's being a point of research in computer science for a bit:, you can read this (it'd put you to sleep quickly): www.cs.umass.edu/lfw/results.html">http://vis-www.cs.umass.edu/lfw/results.html

And when you wake up, you can read up on this:
http://developers.face.com/docs/ (this is "what" facebook uses to match faces, Facebook now owns this, so you know what that means)
http://www.quora.com/Algorithms/What-is-the-algorithm-used-by-Google-Search-by-Image-1 < HOw google approaches the problem which gives an interesting insight on how to tackle this problem.

Anyways, this is not a trivial thing -- so if it's worth your time, well, the links above will explains the magnitude of your "idea".

Sh it! My idea is alreadt open source!

Djtm: Sh it! My idea is alreadt open source!

Why u come resurrect dis thread naaa Nd which of ya idea

Ajibel: Hello people,
I was brainstorming on a 'new' idea with a friend on a boring night we referred to as a 'borinthon' _{(mimicking Facebook's hackathon)} in my room...

See your idea in action at www.tineye.com

This idea ga gan!! Well you fit do am. I guess PIL n Haystack will play a great role in this.

Even if the idea is already in existence, you shouldn't give up. Your own might use a better search algorithm, and might be more effective. I'm certain there were other email hosts before Yahoo and other social networking sites before facebook. So don't give up.

I know a little bit about face detection/face recognition. You can learn/practice using freely available online tutorials. Even if you don't end up doing your 'foogle', you would have added more knowledge to what you have already. Plus there are loads of other applications for recognition/detection algorithms. I've used them for plate number detection systems (traffic cameras use this), etc and security cameras also use face detection/recognition systems to help police to search for criminals. I can give loads and loads of applications.

So basically for a detection system, you want to first identify which part of the image is a 'face'. May sound easy but needs expertise to implement but trust me, people wey dey do am no get two heads so you can. To do this, we do what is called feature extraction. It's 'simply' looking for features which match what a face should look like.

If you know a bit of Machine Learning, then you will be able to train a system to identify faces. There are many free face databases online to use for training and in fact, I believe there are freely available algorithms too. Internet has made life easy

Anyways, start with your programming (codeacademy.com is a good place to learn) and side by side, follow free online face detection tutorials (use google). And if you want to go fully into ML, better start thinking of postgrad study (if you're undergrad already). For starters, try and follow Prof Ng's Machine Learning class on www.coursera.org

Sorry for the long epistle

proudlysports: Even if the idea is already in existence, you shouldn't give up. Your own might use a better search algorithm, and might be more effective. I'm certain there were other email hosts before Yahoo and other social networking sites before facebook. So don't give up.

. . .

Sorry for the long epistle

No need to apologize my man. If only half of programmers thought this way, the world will be a better place.

@Mitey, MySpace started out before FB, which is at the top nau? I havent checked the site yet, buh if i'm to work on this, i can do sth berra
@Picomon, i may take it up later
@Proudlysports, thanks for the epistle and words of encouragement. When i posted this last year, a NLer PM'ed me to assist so i was to work on the backend and dole out new ideas while he was to write the algorithms and backend. We did this for 2weeks by NOV then he disappeard literally. I wasnt inspired so just dumped the whole thing buh as you've offered, i could work on it just to gain knowledge and experience. That's the only reason i'd go back to this cuz i promised myself not to work on this again. Thanks a lot