Background

The ReID-Wild project was initiated in 2025 as a collaborative effort led by researchers committed to advancing wildlife monitoring and conservation technologies. Bringing together a multidisciplinary team spanning computer vision, ecology, and biodiversity management, the project focuses on developing robust, scalable tools for individual stoat identification in real-world environments.

The programme leverages state-of-the-art machine learning methods and close partnerships with conservation organisations to support predator management and long-term ecological monitoring. Building on prior investment in field-camera infrastructure and ecological datasets, ReID-Wild aims to enable accurate, automated, and ethical identification of individual stoats, ultimately supporting data-driven strategies to protect vulnerable

Objectives

A central aim of ReID-Wild is to build an integrated, open, and ethically-grounded data resource that supports the accurate identification of individual stoats in real-world environments. By combining ecological expertise with advanced computer vision techniques, the project seeks to provide a conservation-focused, field-ready tool that strengthens predator-management efforts across Aotearoa New Zealand.

Supporting conservation practitioners with reliable, automated identification will enable more responsive, evidence-based decisions. By improving our ability to monitor stoat populations and their behaviour, the project contributes to more effective protection of native wildlife and habitats, ultimately benefiting biodiversity as a whole.

Context-relevant ecological insights generated through ReID-Wild will guide the design and delivery of predator-control strategies that target the areas of greatest need. This ensures that resources are used efficiently, interventions are better timed, and conservation efforts become more adaptive to changing environmental conditions—helping to close the persistent gap between conservation goals and operational capacity.

Understanding the behaviour and spatial movement of individual stoats can reveal the factors that either hinder or enhance current management approaches. By creating a system for collecting, interpreting, and sharing these patterns in a safe and transparent way, ReID-Wild aims to provide conservation agencies with a clearer picture of what works, what does not, and why.

High-quality, individual-level ecological data has rarely been available at this scale. By enabling non-invasive, automated re-identification across thousands of camera-trap images, ReID-Wild supports long-term monitoring efforts in New Zealand and internationally. The resulting knowledge infrastructure opens new possibilities for cross-ecosystem comparison, more effective predator-control strategies, and accelerated progress toward restoring biodiversity in an era of increasing environmental pressure.