How would a video work? It's something that plays from beginning to end, regardless of whether or not your character is moving.
You don't technically have to add an actual background image to a scene, but you do need to add a transparent png or webp image that has the same dimensions as you want the scene to be as the width & height of the image you assign as the scene background image is what the engine uses to automatically determine the width & height of the scene. Taking what I just said into consideration, you could create scene objects & assign the different background perspective images to those. Create a single value & link it to each of the scene objects containing the background images & then all you need to do is change the number belonging to the value to specify which scene object image should be active/visible.
While what you are wanting to do is technically possible, I don't think you are taking into consideration how much work is involved when it's comes to third person 2D games that contain multiple camera perspectives, camera angles, &/or zoom levels.