Musical Pattern Recognition
Purdue University researchers have developed a musical pattern recognition (MPR) system that accomplishes machine recognition of musical patterns. Taking into account perceptual measures of error, it allows one to find a musical pattern in a collection (database) of musical files. The recognition is robust even if the musical pattern is memorized or recalled incorrectly by a human being. Another potential application is quantifying similarity of two musical compositions (or patterns). Many court cases have dealt with musical plagiarism, relying on human judgment to decide whether one composition or some part of it is similar to another. Having an objective measure of similarity between two musical compositions would be of value in such cases. Finally, music sequencers (software and hardware) can incorporate MPR into their search capabilities, allowing them to perform pattern matching by taking into account perceptual information.
Takes into account the presence of a human being in the systemImproved accuracy and classification of music patterns
Software IndustryComputer Technology
Edward CoylePurdue Electrical and Computer Engineering
United States
None
美国
