作者: Ji Zhang
DOI:
关键词:
摘要: A method for extracting a fingerprint data from video/audio signals, comprising the steps of sending series video frame images into extractor; through and storing database, characterized in that, extractor, said further comprises sub-steps buffer which accepts new image discards oldest first out (FIFO) fashion; sampling by sub-sampler operates on one at time, so that output contains samples obtained multiple while content itself is discarded, after each sampled head buffer, it discarded sampling; making to go divider divides groups frames as be organized segment another one; incorporating optional information formatter form data, formatter, only associated remain; transferring transfer database. According present invention, signals facilitates automatic identification, archiving search without need human visual inspections.