Oh Deer, How Should I Handle This? Seasonal Priors for Selective Wildlife Annotation and Classification

2026-08-03

AI Summary

This study addresses the performance bottleneck in fine-grained classification of red deer in aerial imagery, which arises from sparse target pixels, seasonal appearance variations, and ambiguous annotations. To tackle these challenges, the work introduces a biologically informed soft seasonal prior by explicitly modeling the antler growth cycle of red deer and integrates multimodal data from both RGB and thermal imaging. Through a multi-annotator experiment, the authors demonstrate how seasonality affects annotation quality and model confidence, leading to a novel uncertainty-threshold-based selective prediction mechanism. The proposed approach substantially increases the number of reliably classifiable samples, achieving a coverage accuracy of 98.9%. Notably, the incorporation of the seasonal prior yields particularly pronounced improvements in thermal-image-based classification performance.

Abstract

Fine-grained wildlife classification in aerial imagery is limited not only by model performance, but also by unreliable labels: animals occupy few pixels, key visual cues vary seasonally, and modality-specific evidence can be ambiguous. We study adult-male identification in red deer, where the antler cycle defines predictable windows of reliable evidence for both annotation and prediction. Using 7,295 RGB-only, thermal-only, and matched RGB+thermal crop sets labeled by three annotators, we show that seasonal structure links (I) annotation quality, (II) downstream classification, and (III) selective prediction. Matched RGB+thermal review resolves more samples than either single modality, recovering majority-male labels otherwise missed by RGB or thermal alone, in human based as well as model based classification. Months with high annotator abstention also show lower classifier confidence, and soft seasonal priors mainly benefit the season-limited thermal view. Uncertainty-band abstention further improves covered accuracy up to 98.9%, though at reduced coverage and with deferral that falls disproportionately on males. Overall, a biologically grounded seasonal calendar predicts where annotation and prediction are unreliable, and can guide both annotation protocol design and modality weighting.

Problem

• wildlife classification
• seasonal variation
• annotation reliability
• aerial imagery
• fine-grained recognition

Innovation

• seasonal priors
• selective prediction
• multimodal annotation
• wildlife classification
• uncertainty abstention