METHOD AND SYSTEM FOR AUTOMATIC REAL-TIME FRAME SEGMENTATION OF HIGH RESOLUTION VIDEO STREAMS INTO CONSTITUENT FEATURES AND MODIFICATIONS OF FEATURES IN EACH FRAME TO SIMULTANEOUSLY CREATE MULTIPLE DIFFERENT LINEAR VIEWS FROM SAME VIDEO SOURCE

摘要

A method, a programmed computer system, for example, a network-based hardware device, and a machine readable medium containing a software program for modifying a high definition video data stream in real time, with no visible delays, to add content on a frame by frame basis, thus simultaneously compositing multiple different customized linear views, for purposes such as creating and broadcasting targeted advertising in real time. The method employs conventional video processing technology in novel and inventive ways to achieve the desired objective by passing data selected by the program back and forth between a GPU and a CPU of a computer. The method is also usable with data streams having lower than high definition where real time processing is desired and yields better results than conventional methods. In such applications, all processing may be done by the CPU of a sufficiently powerful general-purpose computer.