We have expanded our Media Intelligence technology to make your Artwork videos even more awesome!
Previously, we used machine learning techniques to analyse your track to get its beats, bars, segments, vocals, drums and more to make responsive music videos. As part of the development of our new artwork video creator, we have expanded our Media Intelligence tech to include Image analysis.
One of the analyses we currently have is predicting the depth of your artwork image using a Deep Learning algorithm.
A large Convolutional Neural Network is used to approximate the depth map of the image - a grayscale image in which darker pixels represent objects closer to the camera. This allows us to create styles which transform your 2D image into 3D space.
To create the 3D effect, a Displacement Map displaces the pixels of your Artwork horizontally and vertically based on the lightness values of the Depth Map.
To try for yourself, click here and you can use artwork styles Depth Beat, Move Your Artwork and Split Glitch.