Film archives around the world contain countless black-and-white films. Since the 1970s there have been attempts to color these afterwards, but the effort is high and the colors often did not match reality. An AI algorithm has been developed at the Technical University of Graz together with a film restoration company from Graz, which is intended to color the old film strips efficiently and realistically.
Historical films often lack that certain something: color. With the help of artificial intelligence, the old shots should be seen in a new color – in realistic colors, as was emphasized. The central requirement is met using trained self-learning neural networks. They can be dynamically influenced by user interaction. The software application was developed by computer scientists led by Thomas Pock of the Institute for Computer Graphics and Visualization at TU Graz together with the Graz-based company HS-Art, which specializes in the restoration of historical films.
Most of it happens automatically
The development that brings color to the past combines deep learning technologies with interactive and automated coloring techniques. The result is an algorithm for a largely automatic but user-controlled inking process. “You always need someone who knows from historical sources what the clothes, facades, etc. looked like at the time. Was the soldier’s uniform green or blue? No algorithm can decide that. But he can learn from it,” says Pock. He leads the Vision, Learning and Optimization (VLO) group at the Graz University of Technology.
The algorithm must be fed with a large collection of training patterns in order to automatically take over the coloring of historical films: “It is about coloring the films as efficiently as possible with as little user input as possible. dictates color for a frame of film and the software then takes over the coloring of other frames,’ explains Pock.
The researchers have analyzed different, and in some cases new, approaches to automated coloring based on artificial intelligence. Then, together with the developers of HS-Art, they implemented the most efficient approach in a prototype application and generated a collection of training examples. Then the implementation of the human-driven control took place to get authentic and appropriate color schemes.
AI tool also repairs damage
With the developed algorithms, the foils can not only be colored, but also scratches, dust and water stains and other damage can be removed. However, this extremely clean restoration is not always desirable. “With historical shots and cinema in general, you need a certain amount of noise, the so-called ‘film grain’, otherwise it doesn’t come across as authentic to the audience. Therefore, after the restoration and staining, the software can artificially generate and add this noise,” emphasizes Pock.
The core algorithm has already been published at an international conference and the source code is freely available. However, software based on this is required for efficient use. This was developed by the project partner HS-Art and is already part of their product portfolio: The so-called “Diamond Film Colorizer” was used, for example, in the ZDFzeit documentary series “Hitler’s Power” to color historical recordings faithfully original.
Source: Krone
I am an experienced and passionate journalist with a strong track record in news website reporting. I specialize in technology coverage, breaking stories on the latest developments and trends from around the world. Working for Today Times Live has given me the opportunity to write thought-provoking pieces that have caught the attention of many readers.
