Descriptor for a video sequence and image retrieval system using said descriptor

A descriptor, embodied within a computer-readable medium, said descriptor being configured to represent, from a video indexing viewpoint, motions of a camera or any kind of observer or observing device, said descriptor being configured for flexibility with respect to how many frames are to be retrieved from a database, and, in particular, to cover the case of one frame, configured to represent the motions within any one frame of a video sequence, and, to cover the case of more than one frame, configured to represent any series comprised of a plurality of frames of the video sequence, the motions represented in said descriptor comprising, for said any one frame and for said series, respectively, at least one of the following basic motion types:

• fixed, panning, tracking, tilting, booming, zooming, dollying and rolling, or any combination of two or more of these types, said descriptor being further configured such that each of said motions represented, except fixed, is oriented and subdivided, in the representation, into two components that stand for two different directions, magnitudes of the respective components each corresponding to a respective predefined size of displacement, the sizes being corresponding values of a dependent variable that defines a histogram for the descriptor,
• wherein the histogram has an independent variable with values that form pairs, each pair corresponding to a different one of said motion types and is configured according to the equation:
• Ntype of motion=Ntype of motion/N
• wherein the subscript “type of motion” represents the independent variable, N represents the number of the frames that are within a window, and Ntype of motion represents the number of frames within the window that have the one of the motion types represented by said subscript.