Evaluating a tandem human-machine approach to labelling of wildlife in remote camera monitoring
Remote cameras (“trail cameras”) are a popular tool for non-invasive, continuous wildlife monitoring, and as they become more prevalent in wildlife research, machine learning (ML) is increasingly used to automate or accelerate the labor-intensive process of labelling (i.e., tagging) photos. Human-machine hybrid tagging approaches have been shown to greatly increase tagging efficiency (i.e., time to tag a single image). However, those potential increases hinge on the extent to which an ML model makes correct vs. incorrect predictions. We performed an experiment using a ML model that produces bounding boxes around animals, people, and vehicles in remote camera imagery (MegaDetector), to consider the impact of a ML model’s performance on its ability to accelerate human labeling. Six participants tagged trail camera images collected from 12 sites in Vermont and Maine, USA (January-September 2022) using three tagging methods (one with ML bounding box assistance and two without assistance).
Complete Metadata
| accessLevel | public |
|---|---|
| bureauCode |
[
"010:12"
]
|
| contactPoint |
{
"fn": "Laurence Clarfeld",
"@type": "vcard:Contact",
"hasEmail": "mailto:lclarfel@uvm.edu"
}
|
| description | Remote cameras (“trail cameras”) are a popular tool for non-invasive, continuous wildlife monitoring, and as they become more prevalent in wildlife research, machine learning (ML) is increasingly used to automate or accelerate the labor-intensive process of labelling (i.e., tagging) photos. Human-machine hybrid tagging approaches have been shown to greatly increase tagging efficiency (i.e., time to tag a single image). However, those potential increases hinge on the extent to which an ML model makes correct vs. incorrect predictions. We performed an experiment using a ML model that produces bounding boxes around animals, people, and vehicles in remote camera imagery (MegaDetector), to consider the impact of a ML model’s performance on its ability to accelerate human labeling. Six participants tagged trail camera images collected from 12 sites in Vermont and Maine, USA (January-September 2022) using three tagging methods (one with ML bounding box assistance and two without assistance). |
| distribution |
[
{
"@type": "dcat:Distribution",
"title": "Digital Data",
"format": "XML",
"accessURL": "https://doi.org/10.5066/P9FGUQEZ",
"mediaType": "application/http",
"description": "Landing page for access to the data"
},
{
"@type": "dcat:Distribution",
"title": "Original Metadata",
"format": "XML",
"mediaType": "text/xml",
"description": "The metadata original format",
"downloadURL": "https://data.usgs.gov/datacatalog/metadata/USGS.64da3a38d34ef477cf3edf0e.xml"
}
]
|
| identifier | http://datainventory.doi.gov/id/dataset/USGS_64da3a38d34ef477cf3edf0e |
| keyword |
[
"Maine",
"USGS:64da3a38d34ef477cf3edf0e",
"Vermont",
"biota",
"camera trap",
"data labelling",
"machine learning",
"tagging",
"trail camera",
"wildlife monitoring"
]
|
| modified | 2023-08-30T00:00:00Z |
| publisher |
{
"name": "U.S. Geological Survey",
"@type": "org:Organization"
}
|
| spatial | -73.2458, 42.8115, -67.9834, 47.1000 |
| theme |
[
"Geospatial"
]
|
| title | Evaluating a tandem human-machine approach to labelling of wildlife in remote camera monitoring |
