Nvidia shows deep learning method to convert 30fps image into slow motion video




Researchers at Nvidia have developed a method to use neural networks for the interpolation of video images. This makes it possible to convert a standard recording in for example 30fps into a slow motion video of, for example, 240 or 480fps.

Nvidia writes that it is already possible to record images with a high frame density, but that this is impractical in many situations. With the deep learning method, called Super SloMo, it would still be possible to add a slow motion effect to an existing standard recording, for example of 30fps. The researchers claim that their method can be used to generate an arbitrary number of frames between two existing frames.

In the paper , which is presented this week at a conference in the US, the researchers write that they use a neural network to predict the content of a frame between two other frames. In addition, one network, a cnn , predicts the movement between two frames, while a second network is used to reduce artefacts.

The training of their model took place on the basis of approximately 11,000 YouTube videos at 240fps, with the model having to predict seven intermediate frames. The researchers state that their method performed better than other existing variants. They show images of the YouTube channel The Slow Mo Guys, which they then delayed further.

Nvidia’s research does not stand alone, so there are other possibilities to delay existing videos later, such as Twixtor.


In: A Technology & Gadgets Asked By: [23167 Red Star Level]

Answer this Question

You must be Logged In to post an Answer.

Not a member yet? Sign Up Now »