A new paper just came out in PNAS that seems to be getting quite a bit of attention - researchers have developed a type of artificial intelligence called "deep neural networks" to automatically identify animals (and more) in images with >93% accuracy. It might be of interest to folks here in our camera trapping community.
http://www.pnas.org/content/early/2018/06/04/1719367115
Abstract
Motion-sensor cameras in natural habitats offer the opportunity to inexpensively and unobtrusively gather vast amounts of data on animals in the wild. A key obstacle to harnessing their potential is the great cost of having humans analyze each image. Here, we demonstrate that a cutting-edge type of artificial intelligence called deep neural networks can automatically extract such invaluable information. For example, we show deep learning can automate animal identification for 99.3% of the 3.2 million-image Snapshot Serengeti dataset while performing at the same 96.6% accuracy of crowdsourced teams of human volunteers. Automatically, accurately, and inexpensively collecting such data could help catalyze the transformation of many fields of ecology, wildlife biology, zoology, conservation biology, and animal behavior into “big data” sciences.
20 July 2018 3:58pm
Thanks, Steph. The camera trap ML nerds among us will have seen drafts of this kicking around in pre-print for a while now. Very cool to see it out finally.
I'll also draw everyone's attention to another study (involving some of the same authors) which has just come out in draft form: "Machine learning to classify animal species in camera trap images: applications in ecology".
Ollie Wearn
Fauna & Flora