No, it's not. Optical flow tells you how far each pixel moved between two images, like between two frames in a video for example. You can use this to determine if objects are moving in a scene. For example if a car or person were walking in front of a non-moving camera, you could find the car or person because only they moved. You could then threshold the optical flow image and get the centroid or area of the moving object, or whatever else you might want. Background subtraction is a different thing. For example if the background scene is a store window, and then the second image has a truck drive in front of it, if you were able to detect the truck (by optical flow or whatever), then subracting a background from it would not make sense - you'd have some weird superposition of a store window and a truck, some kind of ugly blended image. It just wouldn't look right or make sense.
You really have to know your optical and geometric situation to know if it's appropriate to do background subtraction (like fluorescence or radiology) or background division (like most other flat scenes) or if it doesn't make sense to do any kind of background correction at all. For example if you had a face in front of a uniform background with frontal illumination, and another shot of the face with light coming in from 45 degrees from the left, it would not make sense to capture of shot of the background with the 45 degree light and use that to correct the face and expect it to look like the face with the frontal illumination. Because the face is 3D, it just wouldn't work.
