Back in 1989 I was contracted to Kodak to test a system that does basicly
the same thing. A quick check of patent 4,918,588 shows that the Kodak had the same idea pantented in 1990. An operator would
dump documents into a batch scanner and later look at them and associate
text on the image with the image id number. So a user could retrieve a document
image so long as they new at least one word indexed to it. I don't see how Amazon's new system is any different other than OCR replaces a human indexer.
Slashdot Mirror
Back in 1989 I was contracted to Kodak to test a system that does basicly the same thing. A quick check of patent 4,918,588 shows that the Kodak had the same idea pantented in 1990. An operator would dump documents into a batch scanner and later look at them and associate text on the image with the image id number. So a user could retrieve a document image so long as they new at least one word indexed to it. I don't see how Amazon's new system is any different other than OCR replaces a human indexer.