Use of symbol strength and verified watermark detection as basis to improve media-exposure detection
Abstract:
To extract an identifier value from a watermark in media content, a system initially determines the identifier by determining symbol values based on an evaluation of tone strength of each of multiple symbol time segments of the watermark. Further, the system assigns a quality level to the watermark, based on a count of errors in the determined symbol values in view of one or more watermark-structure rules. Still further, if the quality level is threshold low, then the system engages in an error correction process that takes into account, for each symbol time segment whose symbol value may be erroneous, whether that symbol time segment has a threshold strong tone combination that matches the tone combination of a corresponding symbol time segment in a verified watermark. And if the error correction process is successful, the system replaces the determined identifier value with the identifier value of the verified watermark.
Information query
Patent Agency Ranking
0/0