Iterative Human and Automated Identification of Wildlife Images
This work reduces annotation burden for wildlife monitoring researchers by enabling efficient, dynamic model updates, though it is incremental as it builds on hybrid human-machine approaches.
The paper tackles the problem of automating wildlife image recognition from camera traps by addressing the need for large static datasets and long-tailed distributions, achieving ~90% accuracy with only ~20% of the human annotations required by existing methods.
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 iterative human and automated identification approach is capable of learning from wildlife imagery data with a long-tailed distribution. Additionally, it includes self-updating learning that facilitates capturing the community dynamics of rapidly changing natural systems. Extensive experiments show that our approach can achieve a ~90% accuracy employing only ~20% of the human annotations of existing approaches. Our synergistic collaboration of humans and machines transforms deep learning from a relatively inefficient post-annotation tool to a collaborative on-going annotation tool that vastly relieves the burden of human annotation and enables efficient and constant model updates.