Google Is Using AI To Digitize 5 Million Historical Photos (theverge.com)

Google is working with The New York Times to digitize its huge collection of about 7 million historic images. The pictures are apparently kept in the newspaper's "morgue," which contains pictures going back to the 19th century, many of which exist nowhere else in the world. The Verge reports: That's why the company has hired Google, which will use its machine vision smarts to not only scan the hand- and type-written notes attached to each image, but categorize the semantic information they contain (linking data like locations and dates). Google says the Times will also be able to use its object recognition tools to extract even more information from the photos, making them easier to catalog and resurface for future use.

Comments

Dupe

[dereference]

From earlier today no less.

Dupe Dupe Dupe

[Stonent1]

Dupe of URL Dupe dupe dupe of URL.

Re:After digitization, coorelation and AI scanning

[jd]

Silver nitrate and the original magnetic core memory had about the same lifespan. However, you run into problems of size, speed of access, etc. For those youngsters who never encountered it, core memory was an improvement on the Williams memory device that had been used in previous generations of machines. It used slowly-decaying magnetic fields to store information for 100+ years. This made it the world's first electronic non-volatile storage. At a density of 32 kbits per cubic foot, it was also very inefficient.

So, whilst it's not quite correct to say that old-fashioned film is better than all digital in terms of longevity, it's better than most and those that come close in longevity don't come close in information density. High-grade medium film carries a LOT of information in a couple of square inches.

I, too, have an archive.

[jd]

My collection is a few tens of thousands of photographs from three families dating from 1880 or so to 1980. A proper scan that gets out the greatest amount of actual information has yielded an average of about a gigabyte per photograph, so far. That's a lot of information. So my pathetically small collection holds about ten or so terabytes of data.

I've absolutely no idea how I'm going to store that kind of volume of data, it's not like Google will offer.

But I bet you a dozen doughnuts that there are thousands of families in the same boat, who have vast collections of negatives that they'll destroy because they don't have room and don't see an immediate value in.

I also bet that if those thousands of families could be persuaded to get those images scanned, if they'd be willing to contribute to the cost of the collating and storage costs, that it would seriously change the way the past is seen by historians, family history fans and archivists.

And I'd bet that such an archive would have a profoundly greater impact than the NYT archive would.

It won't happen because those aforementioned families will be stubborn and prefer destruction over conservation, because none of the cloud vendors would be interested in helping disseminate information of individually uncertain value, and because most people see history as someone else's problem.