Gaspard Dussert

This paper improves camera trap classification models by incorporating self-attention to leverage information from multiple animal detections, enhancing species classification accuracy in both mono- and multi-species scenarios.

This study uses vision-language foundation models to classify animal behaviors in camera trap images without prior training, achieving over 91% accuracy in three ungulate species.

This study evaluates the calibration of deep learning models for camera trap image sequences, showing that temperature scaling improves confidence score interpretability for ecological applications.