Camera trap classification with deep learning under ground truth uncertainty
2026-08-31 • Computer Vision and Pattern Recognition
Computer Vision and Pattern Recognition
AI summaryⓘ
The authors studied how uncertain labels from volunteer citizen scientists affect training deep learning models on ecological images. They found that including some disagreement in training labels actually improved the model’s accuracy, especially for images that volunteers found hard to classify. Using pre-trained models on general image datasets helped too, making training faster and better. However, these improvements didn’t always transfer well to other datasets. Their work suggests that embracing label uncertainty and using pre-trained models can help combine human and machine classifications more effectively.
supervised deep learningcitizen sciencecamera trap imagesground truth uncertaintypre-trainingImageNetspecies classificationlabel disagreementclass imbalancemodel generalization
Authors
Leonard Hockerts, Peter S. Stewart, Sarthak Arora, Tiffany J. Vlaar
Abstract
Supervised deep learning methods enable the rapid processing of ecological image data, but depend on a costly annotation process. Consequently, training labels are commonly derived from volunteer citizen science projects. However, disagreement among volunteers introduces uncertainty in the "ground truth" data that are assumed to be correct for model training and validation. Using two datasets containing camera trap images with associated volunteer and expert classifications, we investigated the effects of training under higher ground truth uncertainty. We observed improved overall test accuracy, particularly for images that were more difficult for volunteers. Species-level accuracy also generally improved, but generalisation to a different dataset did not. The benefits of ground truth uncertainty were enhanced by pre-training on ImageNet. Pre-training also reduced the number of training epochs required; further reductions in computational cost, but not gains in accuracy, resulted from additional pre-training on other camera trap images. With unbalanced training data, we still observed a clear benefit of increased ground truth uncertainty for overall accuracy, especially on difficult images. Class imbalance improved accuracy for common species, reduced rare species accuracy, and changed patterns of misclassification to more closely resemble mistakes made by volunteers. Our findings have implications for applying deep learning across ecological image types with multiple labels. Practitioners can improve accuracy, especially on difficult examples, by including moderate levels of label disagreement during training and using models pre-trained on general image data. In addition to improving the use of citizen science-derived labels in model training, our study suggests avenues for more effectively integrating human and deep learning classifications in combined workflows. (abridged)