What is the difference between videopose's 3d skeleton point and kinetic`s skeleon

[LynCoding]

I want to convert the prediction result of videopose into coordinate data in kinetic format.In videopose prediction results, i know that x ,y represent the two-dimensional coordinate position of the corresponding skeleton point,maybe,they have been normalized.but,i dont know what the z axis represents.