A video camera captures multiple images each second (often 30 or 60). Each image is called a frame. In raw formats these frames are stored one by one, but more comoact formats, such as mp4, use compression so that only frame-to-frame changes are stored. Some computer vision techniques work by analying each frame separately and then looking at the way these connect, but others make use of the moving image, for example calculating optical flow, which can help separate background from foreground objects.
Used on Chap. 19: page 470