Old videos are not of the best quality, they have a lot of noise and ambiguity, and with all that, they have no color.
Luckily, artificial intelligence has advanced enough that repairing those videos is now possible.
Interesting: try the free AI tool that very accurately colors black and white images.
The paper dealing with this problem shows how AI successfully restores contrast, and clarity and removes flicker from old videos.
On top of all that, with just 6 reference images, the AI adds color to the video almost without a mistake!
In the video above, in the upper left corner, there is the original video, and in the very lower right corner, there is an improved video.
Satoshi Iizuka and Edgar Simo-Serra, co-authors of this paper, explained that this model (Temporal Neural Network) aims to perform multiple operations on the video:
Remastering old movies involve a variety of sub-tasks including super-resolution, noise reduction, and contrast correction, which restores worn-out images to their original appearance. In addition, due to technological limitations at the time, most films were shot either in black and white or in very poor color, which is why colorization is necessary.
So far, this technique has not found commercial application, but it would be entirely expected that large film industries, such as Hollywood, will use AI as an aid soon.
Comments
Post a Comment