Improving Wildlife Out-of-Distribution Detection: Africas Big Five

2025-06-07

AI Summary

To address the overconfidence of closed-world classification models on out-of-distribution (OOD) samples in wildlife recognition—particularly for early warning of human-wildlife conflict—this work focuses on OOD detection for Africa’s “Big Five” (lion, leopard, elephant, buffalo, rhinoceros) under open-world conditions. We propose a lightweight, pre-trained feature-driven OOD detection framework that uniquely integrates Nearest-Class-Mean (NCM) scoring with non-parametric contrastive learning. Evaluated systematically against multiple state-of-the-art methods on a unified benchmark, our NCM-based approach achieves new best results: +2% in AUPR-IN, +4% in AUPR-OUT, and +22% in AUTC over the strongest baseline, demonstrating significantly improved cross-threshold generalization. The code and a dedicated benchmark dataset are publicly released.

Technology Category

• Machine Learning: Multi-class/Multi-label Learning & Extreme Classification
• Computer Vision: Object Detection & Categorization
• Data Mining & Knowledge Management: Anomaly/Outlier Detection

Application Category

• Web Mining and Content Analysis: Large pretrained models with web data
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

Mitigating human-wildlife conflict seeks to resolve unwanted encounters between these parties. Computer Vision provides a solution to identifying individuals that might escalate into conflict, such as members of the Big Five African animals. However, environments often contain several varied species. The current state-of-the-art animal classification models are trained under a closed-world assumption. They almost always remain overconfident in their predictions even when presented with unknown classes. This study investigates out-of-distribution (OOD) detection of wildlife, specifically the Big Five. To this end, we select a parametric Nearest Class Mean (NCM) and a non-parametric contrastive learning approach as baselines to take advantage of pretrained and projected features from popular classification encoders. Moreover, we compare our baselines to various common OOD methods in the literature. The results show feature-based methods reflect stronger generalisation capability across varying classification thresholds. Specifically, NCM with ImageNet pre-trained features achieves a 2%, 4% and 22% improvement on AUPR-IN, AUPR-OUT and AUTC over the best OOD methods, respectively. The code can be found here https://github.com/pxpana/BIG5OOD

Problem

• Detects out-of-distribution wildlife species accurately
• Improves human-wildlife conflict mitigation using Computer Vision
• Evaluates feature-based methods for better generalization in OOD detection

Innovation

• Uses parametric Nearest Class Mean method
• Applies non-parametric contrastive learning approach
• Leverages ImageNet pre-trained features effectively