Alan Trinder asks:
I have a number of historical photos and there are some people that I cannot identify. This must be the case for many people looking into their family history or old photos. I would like to find a database for photos that offers search functions by specific fields including inbuilt face recognition.
If possible I would like to make this a public database (with membership etc) so it can be used as a research tool.
7 Answers -
Jay.0
February 19, 2012Google image search : http://www.google.com/imghp
and tineye : http://www.tineye.com/
can help, to find who they are.
If these people are famous or featured on some popular websites.
Karkala Nayak
February 19, 2012Hello Alan,
There is a website which lets you search for images on the web using the face recognition
here is the link::
http://www.pictriev.com/facedb/fs2.php
Also a website that lets you build an app that is equipped with facial recognition
check face.com…Facebook uses their technology for facial recognition..that means this website can help you…
http://face.com/
You can check out face.com’s API for developers.. here is the link…
http://developers.face.com/
Here is the list of Similarity based image search engines list..
http://www.searchenginejournal.com/7-similarity-based-image-search-engines/8265/
Also, you can check this site here::
http://alipr.com/
I hope it helped::
Hi Karkala, your sites very helpful for what i am looking to do
February 20, 2012Thanks for the feedback Alan. Glad to know that the answers helped…
February 20, 2012Good Day..
Anonymous
February 19, 20123 Fascinating Search Engines That Search For Faces
http://www.makeuseof.com/tag/3-fascinating-search-engines-search-faces/
Top 10 reverse image search engines
http://www.comptalks.com/top-10-reverse-image-search-engines/
Alan Trinder
February 20, 2012many thanks for some superb suggestions
M-t
February 20, 2012Ridiculous idea. Would take a massive herculean effort to get almost nowhere with that and then it would be completely unreliable for an multitude of reasons such the fact the age, light, angles, size, quality and all sorts of things would make the data nearly useless.
I kind of like ridiculous ideas, manned travel to the moon in the 1960s, being able to fly, the wheel, computing power (how quickly does that enlarge). We just have a different outlook, you and I
February 21, 2012James Bruce
February 20, 2012None of these tools will do what you’re asking for, since the task is impossible. As cool as this would be, there’s simply no way to do facial recognition on grainy, faded old photographs.
Ultimately, this is task for which the human brain is much more suited, in terms of structure as well as processing power. Now ask yourself – would *you* be able to match two faces from old grainy photographs? I very much doubt it to be honest. So even if computers were that powerful, this would still be impossible.
Hi James
February 21, 2012perhaps the way I describe this over emphasise the reliance on the face recognition. Thanks for your comments will certainly help me keep things in perspective.
K W H
February 20, 2012Yes but no. The trick is that facial recognition, at least on the consumer level, is overrated. Two big examples of available recognition technology are iPhoto and Facebook’s auto-tagging features. This form of recognition only ties together similar looking faces (same lighting, angle and features) and prompts the user to name or tag the face. It is just not that clever at recognizing the same face from a new angle, and more importantly it only appears to be successful because it works with a limited “corpus” to match against. Example: you might have 30 people tops tagged in iPhoto. Facial recognition doesn’t have to figure out who the person is, just make a 1 in 30 guess with reasonable confidence. Same with Facebook when you have 50-100 friends.
The problem is that this doesn’t work large scale – take a million photos and 10 million different potential people it could be. Tons of false positives that a human has to figure out and tag, not to mention the algorithm has to be more discriminating – it doesn’t just have to tell your kid apart from grandpa, but many hundreds of people who might be facially nearly identical. It’s true that things like Google Image Search and Tineye can be used to find an almost identical photo, but when it comes to matching people it’s the human identification factor which is most important. A genealogical database like you describe is only going to be as useful as the info put in, face recognition isn’t a silver bullet in this case.
Hi KWH
Many thanks for your comments, my description of my objectives perhaps overemphasised the face recognition. Your input very valuable many thanks
February 21, 2012