Open-weight models enable local wildlife species identification with modest data
Standing on Open Shoulders: Regional Fine-Tuning of Open-Weight Camera-Trap Models for UK Wildlife with Modest Data and Compute
Computer Vision and Pattern Recognition
Summary
Identifying specific animal species in local regions, like different squirrels in the UK, is hard with big AI models trained elsewhere. The authors show that small teams using open AI models can fine-tune these models with a small amount of local camera-trap images and a modest GPU. Their fine-tuned model accurately distinguishes 28 UK wildlife classes, including tricky pairs like native and invasive squirrels. This makes regional animal monitoring more accessible but still depends on major groups releasing their AI models openly.
What this means in practice
- •For wildlife conservation teams: Build region-specific animal species classifiers from limited local images using open AI models and modest GPU resources.
- •For environmental monitoring services: Offer improved automated camera-trap image analysis tailored to local species differences for more accurate wildlife surveys.$Commercial implications: Enables selling enhanced wildlife monitoring tools with better local species recognition to environmental agencies.
Authors
Paul Fergus
Abstract
Camera-trap species classification is dominated by large models from well-resourced organisations, which often lack the species-level distinctions regional conservation needs, such as native red versus invasive grey squirrels in the UK. We ask whether open weights let a small team with a modest dataset and a single GPU take a real part in AI development. We fine-tuned the open DeepFaune classifier on 28 UK classes, paired it with the unmodified MegaDetector v6, and compared the resulting DeepFaune-UK v1.2 with DeepFaune v1.5.0 and SpeciesNet on 20,212 hand-verified MammalWeb citizen-science images, after removing sequences sharing any of 1,039 images with our training archive. DeepFaune-UK achieved a strict image-level accuracy of 88.8%, including empty frames, against 56.5% and 59.0% for the baselines, but at DeepFaune's coarser label resolution DeepFaune v1.5.0 was more accurate, though not significantly under contributor resampling (91.4% vs. 88.8%). As DeepFaune v1.5.0 uses a newer DINOv3-based classifier, we applied the identical recipe, without new data, to that classifier. The resulting DeepFaune-UK v1.3 reached 90.6% strict accuracy, narrowed the like-for-like gap to 0.8 points and reduced roe deer misclassified as fallow deer by 69%. On 1,850 images from the 26 contributors with no overlapping images, the three DeepFaune-based models did not differ significantly (90.5%, 90.6%, 89.9% like-for-like). Fine-tuning on 10% of the training crops (about 150 per class) gave out-of-distribution accuracy similar to using all of them, and a DINOv2 run on this subset took about 22 minutes with under 4 GB of GPU memory. Open weights thus make regional model development feasible for small teams, but the largest like-for-like gain came from a newer base model trained by others, so this democratisation depends on large actors continuing to release open weights, under licences that may change.