Baseband video is the DC to about 6 MHz signal that encodes the brightness channel, and sync. If it includes colour too then it's called Composite video.
Brightness and sync are truly baseband, a specific voltage range represents the black to white brightness over time as the beam scans the display, with sync represented by pulses to a different voltage. Colour is actually modulated onto a carrier, even at 'baseband'.
In the diagram you quote, the baseband is effectively the signal just after the detector block. In a receiver+display there's no need to generate a standards-compliant video signal.
IF, as you correctly surmise, contains the baseband video but modulated to some intermediate, frequency, perhaps about 50 MHz, suitable for further filtering and amplification before detection. It is solely a choice of the receiver designer, and is invisible from the outside. The receiver you're thinking of, takes in RF and produces baseband video.