Dimitris Poulopoulos

Machine Learning Engineer | Researcher

The Slow Mo Guys now Matched by AI

11 October 2022

We can all agree that slow-motion videos create this dramatic effect, which certainly adds an extra flavor to specific scenes. On the other hand, slowing down a video might create several unwanted artifacts, which could deem it unwatchable. But what makes a slow-motion video great?

Among other things, I would argue that the most important aspect is its FPS count: how many frames per second are rendered to the screen. Too low and the video stutters, creating an unpleasant viewing experience. Thus, we need to have an expensive camera able to record at a high frame rate. Or, we could use Deep Learning and try to synthesize intermediate images between two consecutive frames. And did I mention that this will cost us 0 more dollars?

Super slow-motion is just one of the applications of Video Frame Interpolation or VFI as it’s commonly known in the field of Computer Vision. This story presents RIFE: Real-Time Intermediate Flow Estimation for Video Frame Interpolation. We touch on what it is bringing to the table and take it for a quick test drive using our own videos. Let’s dive in!

Continue reading ->