Attempting to create an automatic detection algorithm to scan large data sets. I want to pull out beluga vocalizations to determine spatio-temporal habitat use from several years of hydrophone recordings.
Our foundation has been funding for several years a project which has developed an automatic detection system. This has been designed to detect, extract and store in a database killer whale vocalizations but theoretically it should work with any sound event. It would be great to have one of your recordings to make a test and, if if works, we can find the way to process your recordings, as the system stills under development and there is no a version for distribution yet.