Review: Content-Based Image Retrieval: Tools, Writings and Demos
We've by writing and posting about content-based Image Retrieval for years on ResourceShelf and also Search Engine Watch. In recent years we've added the related (and soon to be mainstream) idea of cameraphone searching to the mix. Here are links to several of the posts.
While many have recently been introduced to "face searching" at Exalead and Google, it has been available elsewhere for some time. Freenet.de is an example of a site that continues to offer face search and more using technology from Cobion (now part of IBM, Cobion itself was acquired by ISS in 2004, ISS by IBM in 2006).
And More? Freenet.de also offer OCR image search (find a word(s) "inside" an image(s)).
Here's the image search UI (in German). The first option below the search box (if checked/ticked) requires that the word(s) in the search box must be found in the image. The third box (if checked/ticked) requires a face (of a person) in the image. Here's a screen cap of a search that combines both criteria. The search was for Lufthansa.
Another demo we've written about for years is from LTU where you can find related images based on both assigned keywords and what they call the "image DNA." Btw, LTU powers "face search" on Exalead.
Of course we cannot for QBIC? This content-based tool from IBM has been around a very long time in Internet years and is still being used on The State Hermitage Museum site.
Today, a note about Airliners.net. This is very well-known source to find pictures of airplanes, airports, etc. and we've mentioned it in the past.
What ResourceShelf wants to point out is that the Airliners.net database contains not only plenty of images (over one million) but also a lot of metadata that can help users build very precise queries. Here's the search interface and a search for:
+ Boeing 777-200 aircraft with British Airways markings
+ Taken in 2006
+ At or near London Heathrow
Yes, another example of the power of good metadata. Though it's behind the scences, is Pandora/Music Genome Project. this music database is another powerful example of great metadata. Where does it come from? How do they get it? More here.
OK, fine. But what does any of this have to do with content-based image retrieval?
Have fun. As you can tell, Content-Based Image Retrieval and cameraphone searching our two area we monitor closely. Stay tuned for more. It's coming fast and furious.
A family of resources to help information workers be more effective, raise the value of information in their organisations and contribute to success. Read more »
Recently I have found myself cooing over visualisation maps (and heat maps) of health and well being resources. The content rich data is overlayed with mapping technologies, and some interesting themes and patterns are emerging.
A lot of the talk around social media in the last year has been around information overload. Social media has provided us with new and exciting ways to create content. But it has also meant learning new ways to manage and engage with social media tools. Are we teetering on the edge of an information overload precipice?
Information overload is a figment of your imagination. Or a failure of your filter. Or a symptom of your technological submissiveness. Depends on who you ask.
What if you had to sort through 3.5 million articles and social media posts a day and try to pull out the most relevant items for your organisation? What if you then had to cobble it all together into something readable for your top groups and executives in your organisation?
Alacra Compliance saves time by aggregating information from both free and fee-based sources and enabling users to conduct an accurate federated search across these sources (coined “simultaneous search” by Alacra).
FreePint Family
Quilting big data threads
The fallacy of information overload
Information overload: fact, fantasy or filter failure? 
