Use of object detection in camera trap image identification: assessing a method to rapidly and accurately classify human and animal detections for research and application in recreation ecology
Bibliographic record
Abstract
Abstract Camera traps are increasingly used to answer complex ecological questions. However, the rapidly growing number of images collected presents technical challenges. Each image must be classified to extract data, requiring significant labour, and potentially creating an information bottleneck. We applied an object-detection model (MegaDetector) to camera trap data from a study of recreation ecology in British Columbia, Canada. We tested its performance in detecting humans and animals relative to manual image classifications, and assessed efficiency by comparing the time required for manual classification versus a modified workflow integrating object-detection with manual classification. We also evaluated the reliability of using MegaDetector to create an index of human activity for application to the study of recreation impacts to wildlife. In our application, MegaDetector detected human and animal images with 99% and 82% precision, and 95% and 92% recall respectively, at a confidence threshold of 90%. The overall time required to process the dataset was reduced by over 500%, and the manual processing component was reduced by 840%. The index of human detection events from MegaDetector matched the output from manual classification, with a mean 0.45% difference in estimated human detections across site-weeks. Our test of an open-source object-detection model showed it performed well in partially classifying a camera trap dataset, significantly increasing processing efficiency. We suggest that this tool could be integrated into existing camera trap workflows to accelerate research and application by alleviating data bottlenecks, particularly for surveys processing large volumes of human images. We also show how the model and workflow can be used to anonymize human images prior to classification, protecting individual privacy. Impact Statement We developed and tested a workflow for classifying camera trap images that integrated an existing object-detection model with manual image classification. Our workflow demonstrates an increase in efficiency of 500% over manual labelling, and additionally includes a method to anonymize human images prior to archiving and classification. We provide an example of the application of these tools to ease data processing, particularly for studies focused on recreation ecology which record high volumes of human images. Data lags due to processing delays have the potential to result in sub-optimal conservation decisions, which may be alleviated by accelerated processing. To our knowledge, this is the first in-depth assessment of the practical application of such technology to real world workflows focused on human detections.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.005 | 0.009 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.003 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.002 | 0.001 |
| Open science | 0.002 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".