Big data comes in all forms and corporations worldwide are leveraging the power of machine learning to make decisions, automate processes, and enhance their offerings. The adult industry is no different with Pornhub’s announcement of voluptuous Victorians and silent seductors now released to a modern audience in full colour, despite being originally aged black and white footage.

(Note: the links below are not safe for work, in case you had any doubts ...)

Pornhub announces its audience can stop rewatching the same old scenes and instead transport themselves to a bygone era. Of course, for data scientists perhaps you may need to inspect the catalogue for yourselves “for science.”

Pornhub has titled the project "Remastered," delivering a modern restoration of some of the oldest and steamiest erotic films ever produced. The library includes 20 scenes ranging from the 1890s through the 1940s and features the work of Thomas Edison among others.

“Since the moment someone figured out how to record moving images, people have been making adult movies,” said Pornhub. “We thought it was important to not only preserve but modernise these films, and we are very excited to include this selection of artfully restored vintage material to our library, allowing users to indulge in old school erotica from a bygone era.”

Re-colourising films is not new, but previously was a tedious manual process or employed algorithms unsuitable for Pornhub’s project because they were trained with millions of, shall we say, “safe for work” imagery.

By introducing artificial intelligence machine learning to Pornhub's massive content library - let’s hope it was a proper introduction - the algorithms learned a naughty new trick, allowing it to detect certain body parts - “the subtle shades of penises, nipples, and turn-of-the-century bush,” in Pornhub’s words. iTWire is also told the plots, such as they are, are time-honoured favourite categories.

For those who appreciate the technology - the digital equivalent of "reading the articles," perhaps, the process leverages algorithms to prepare images, reduce noise, and sharpen and contrast, to colourise using deep learning, boosting the film to 60 frames per second, rescaling to 4K resolution, and digitally remastering to clean artefacts, stabilise the video, and reduce flicker, then finally to remaster the audio or add a new audio track completely.

To train the machine learning algorithm Pornhub paired black and white images along with the colourised version out of its massive library of full-colour pornography. This allowed the algorithm to then generalise its findings and colourise images that had previously never had colour.

You can have a sneak preview of the collection in this definitely safe for work video below: