Location Invariant Animal Recognition UsingMixed Source Datasets and Deep Learning
A. J. Shepley, D. G. Falzon, P. Meek, P. Kwan
This paper trains a deep learning model using FlickR images to improve camera trap object detection, achieving higher accuracy and location invariance compared to models trained on single-location datasets.
bioRxiv · May 13, 2020 · PreprintEcology