Being confident in confidence scores: calibration in deep learning models for camera trap image sequences
Open Access
- 16 June 2024
- journal article
- research article
- Published by Wiley in Remote Sensing in Ecology and Conservation
- Vol. 11 (1) , 88-99
Abstract
In ecological studies, machine learning models are increasingly being used for the automatic processing of camera trap images. Although this automation facilitates and accelerates the identification step, the results of these models may lack interpretability and their immediate applicability to ecological downstream tasks (e.g. occupancy estimation) remains questionable. In particular, little is known about their calibration, a property that allows confidence scores to be interpreted as probabilities that model's predictions are true. Using a large and diverse European camera trap dataset, we investigate whether deep learning models for species classification in camera trap images are well calibrated. Additionally, as camera traps are often configured to take multiple photos of the same event, we also explore the calibration of predictions aggregated across sequences of images. Finally, we study the effect and the practicality of a post‐hoc calibration method, i.e. temperature scaling, for predictions made at image and sequence levels. Based on five established models and three independent test sets, we show that averaging the logits over the sequence, selecting an appropriate architecture, and optionally using temperature scaling can produce well‐calibrated models. Our findings have clear implication for, for instance, the calculation of error rates or the selection of confidence score thresholds in ecological studies making use of artificial intelligence models.Keywords
All Related Versions
Funding Information
- Grand Équipement National De Calcul Intensif (2022AD010113729)
This publication has 22 references indexed in Scilit:
- Using the multivariate Hawkes process to study interactions between multiple species from camera trap dataEcology, 2024
- Automated visitor and wildlife monitoring with camera traps and machine learningRemote Sensing in Ecology and Conservation, 2023
- Automated wildlife image classification: An active learning tool for ecological applicationsEcological Informatics, 2023
- Efficacy of machine learning image classification for automated occupancy‐based monitoringRemote Sensing in Ecology and Conservation, 2023
- BatNet: a deep learning‐based tool for automated bat species identification from camera trap imagesRemote Sensing in Ecology and Conservation, 2023
- A ConvNet for the 2020s2022
- The effect of urbanization on spatiotemporal interactions between gray foxes and coyotesEcosphere, 2022
- AI Naturalists Might Hold the Key to Unlocking Biodiversity Data in Social Media ImageryPatterns, 2020
- Automatically identifying, counting, and describing wild animals in camera-trap images with deep learningProceedings of the National Academy of Sciences, 2018
- Rethinking the Inception Architecture for Computer Vision2016