Complex computer ideas and old, fantastical stories have a long history together. Neo followed the White Rabbit to learn more about the Matrix, and the exit from the program was itself through a looking-glass. Many malware ideas try to make you ignore the man behind the curtain, and trojans speak for themselves. But deep learning is forging an even closer connection to old stories by repairing old film. Here's how:

1. Color recognition programs can accurately change black and white movies.

There's always been a demand to change beloved black and whites to colorful productions. But the output was always a little lacking. The colors would be muted or not quite right, and each endeavor cost more than was sustainable.

But deep learning programs that can teach themselves to guess colors down to the pixel can help. Instead of requiring manual coloring, or even humans to check and correct progress every few frames, projects get faster and cheaper. Soon all of the old classics might be in HD color.

2. AI is rebuilding fuzzy details.

Color isn't the only thing missing from a lot of old film reels. Water damage and rot have eaten away at the edges of more than one notable recording before they were digitally backed up. Old films also have some fuzzy moments, blurry details, and missing microseconds. While deep learning might not be enough to get AI smart enough to completely fill in lost Doctor Who episodes, algorithms can fill in the gaps and recoup lost edges.

3. Language recognition programs can help with captions.

Old movies have pretty good audio. But movies these days come with a variety of captions, explanatory audio descriptions, and other options to increase accessibility. AI is the key to churning those elements out with greater accuracy and clarity. Because AI learns quickly and costs much less than human transcriptionists, more films can be enjoyed by larger audiences.

Deep learning isn't just improving your cinematic experience. Go to folio1's blog for more news and updates or give our team a call today.