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

哦，鹿啊，我该如何处理？用于选择性野生动物标注与分类的季节性先验

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.

摘要： 航空影像中的细粒度野生动物分类不仅受限于模型性能，还受到不可靠标签的制约：动物在图像中占据的像素极少，关键视觉特征随季节变化，且特定模态的证据可能存在歧义。

We study adult-male identification in red deer, where the antler cycle defines predictable windows of reliable evidence for both annotation and prediction.

我们研究了马鹿（red deer）成年雄性的识别问题，其中鹿角的生长周期为标注和预测提供了可靠证据的可预测窗口。

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.

通过使用由三名标注员标注的 7,295 组仅 RGB、仅热成像以及匹配的 RGB+热成像裁剪数据集，我们证明了季节性结构将以下三者联系起来：(I) 标注质量，(II) 下游分类，以及 (III) 选择性预测。

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.

匹配的 RGB+热成像审查比单一模态能解决更多的样本，在人工分类和模型分类中，都能找回仅靠 RGB 或热成像无法识别的雄性标签。

Months with high annotator abstention also show lower classifier confidence, and soft seasonal priors mainly benefit the season-limited thermal view.

标注员弃标率较高的月份，分类器的置信度也较低；而软季节性先验主要对受季节限制的热成像视图有益。

Uncertainty-band abstention further improves covered accuracy up to 98.9%, though at reduced coverage and with deferral that falls disproportionately on males.

不确定性区间弃标进一步将覆盖准确率提高到了 98.9%，尽管这以降低覆盖率为代价，且推迟处理的情况在雄性样本中更为突出。

Overall, a biologically grounded seasonal calendar predicts where annotation and prediction are unreliable, and can guide both annotation protocol design and modality weighting.

总的来说，基于生物学的季节性日历可以预测标注和预测在哪些情况下不可靠，并能指导标注协议的设计和模态权重的分配。