TY - RPRT
TI - Feasibility Analysis: Using Artificial Intelligence to Match Photographed Lateral Ridges of Gray Whales
AU - Holmberg, J
AU - Parham, J
AU - Blount, A
AB - Wild Me (wildme.org) completed all tasks for BOEM Award 140M0120P0023 and NOAA Award 1305M320PNFFR0479 (collectively the “AI for Gray Whales” project) and is submitting this final report to complete the project. Wild Me evaluated four distinct computer vision approaches to reliably reidentify gray whales (Eschrichtius robustus) from lateral photos.  Among the evaluated techniques, the HotSpotter and PIE algorithms provided the most overall matching power with an additive performance of top-1 rank of 70% and top-12 of 92%, depending on their chosen configuration and the selection of test data. All developed and tested machine learning models and ID algorithms evaluated under these awards are now available in Flukebook.org for evaluation and use.
CY - Camarillo, CA, USA
DA - 2021/08//
PY - 2021
SP - 29
PB - Bureau of Ocean Energy Management (BOEM)
SN - BOEM 2021-059 
UR - https://www.boem.gov/BOEM-2021-059
LA - English
KW - Marine Mammals
KW - Cetaceans
ER -