Neural networks are one of the most exciting developments in the computing world, with a wide range of potential applications. This video shows what happens when a neural network is used to take footage from 1896 and upscale it to 4K, add a faster frame rate, and even add color.
Denis Shiryaev created this video by taking footage from "L’Arrivée d’un train en gare de La Ciotat" (The Arrival of a Train at La Ciotat Station), a short film made in 1895 and premiered in 1896 by the famous Lumière brothers. The film shows a train pulling into the station and people disembarking. The film is famous for the legend that early movie-goers fled the theater in terror at the sight of the approaching train. You can watch the original video above and the enhanced version below.
Shiryaev used Topaz Labs Gigapixel AI to upscale the original 4K, then used DAIN (Depth-Aware Video Frame Interpolation) to add additional frames to raise the rate to 60 fps. You can see a version that has been colorized by DeOldify Neural Network below.
Altogether, it looks to be quite the exciting application of the technology with a lot of future potential!