2.5D video

We used the Azure Kinect sensor to record a three-dimensional video. 
It had the highest resolution of shooting at a depth of 1024x1024. The only limitation was in the frame rate – it could only be done at 15FPS. That was enough for us. 

The DepthKit program was used as the software. It can work with Azure Kinect and simplifies the process of creating such a video:
– Possibility to record data with the basic sensor settings: shooting resolution, white balance adjustment, exposure, etc
– Having a preview in a small editor and the ability to select a depth threshold in order to crop the background or select some objects and crop the whole video. This point is important because when shooting at a resolution of 1024x1024, the video with a depth has the shape of an octagon and does not coincide with the shape of the video shot on an RGB camera, which is rectangular
– Possibility to add a black-and-white mask to the video to highlight a certain object, such as a person

An example of selecting the depth value, and all the examples are taken on Azure Kinect: