Iterative Human and Automated Identification of Wildlife Images

  title={Iterative Human and Automated Identification of Wildlife Images},
  author={Zhongqi Miao and Ziwei Liu and Kaitlyn M. Gaynor and Meredith S Palmer and Stella X. Yu and Wayne M. Getz},
Camera trapping is increasingly used to monitor wildlife, but this technology typically requires extensive data annotation. Recently, deep learning has significantly advanced automatic wildlife recognition. However, current methods are hampered by a dependence on large static data sets when wildlife data is intrinsically dynamic and involves long-tailed distributions. These two drawbacks can be overcome through a hybrid combination of machine learning and humans in the loop. Our proposed… Expand
1 Citations

Figures and Tables from this paper

Deep Long-Tailed Learning: A Survey
  • Yifan Zhang, Bingyi Kang, Bryan Hooi, Shuicheng Yan, Jiashi Feng
  • Computer Science
  • ArXiv
  • 2021
A comprehensive survey on recent advances in deep long-tailed learning is provided, highlighting important applications of deepLongtailed learning and identifying several promising directions for future research. Expand


ImageNet: A large-scale hierarchical image database
A new database called “ImageNet” is introduced, a large-scale ontology of images built upon the backbone of the WordNet structure, much larger in scale and diversity and much more accurate than the current image datasets. Expand
Robust ecological analysis of camera trap data labelled by a machine learning model
This is a list of winners and nominees for the 2016 Rogers Cup, which was won by Britain's Andy Murray in the final in Montreal. Expand
Snapshot Safari: a large-scale collaborative to monitor Africa’s remarkable biodiversity
Volume 117| Number 1/2 January/February 2021 Commentary Snapshot Safari: A large-scale collaborative to monitor Africa’s remarkable biodiversity AUTHORS: LainExpand
What Should Not Be Contrastive in Contrastive Learning
This work introduces a contrastive learning framework which does not require prior knowledge of specific, task-dependent invariances, and learns to capture varying and invariant factors for visual representations by constructing separate embedding spaces, each of which is invariant to all but one augmentation. Expand
"How many images do I need?" Understanding how sample size per class affects deep learning model performance metrics for balanced designs in autonomous wildlife monitoring
The issues of deep learning model performance for progressively increasing per class (species) sample sizes are explored and generalizes additive models (GAM) are shown to be effective in modelling DL performance metrics based on the number of training images per class, tuning scheme and dataset. Expand
A Simple Framework for Contrastive Learning of Visual Representations
It is shown that composition of data augmentations plays a critical role in defining effective predictive tasks, and introducing a learnable nonlinear transformation between the representation and the contrastive loss substantially improves the quality of the learned representations, and contrastive learning benefits from larger batch sizes and more training steps compared to supervised learning. Expand
An empirical evaluation of camera trap study design: How many, how long and when?
ABSTRACT 1. Camera traps deployed in grids or stratified random designs are a well-established survey tool for wildlife but there has been little evaluation of study design parameters. 2. We usedExpand
Born‐digital biodiversity data: Millions and billions | 1 Museum specimens have always provided the most basic informa‐ tion about the spatial distribution of life on earth: which species live where and when. TheseExpand
Energy-based Out-of-distribution Detection
This work proposes a unified framework for OOD detection that uses an energy score, and shows that energy scores better distinguish in- and out-of-distribution samples than the traditional approach using the softmax scores. Expand
Improving the accessibility and transferability of machine learning algorithms for identification of animals in camera trap images: MLWIC2
By including many species from several locations, the species model is potentially applicable to many camera trap studies in North America and the empty-animal model can facilitate removal of images without animals globally. Expand