BeetleVerse: A study on taxonomic classification of ground beetles

2025-04-18

AI Summary

To address key bottlenecks in large-scale automatic ground beetle classification—namely, subtle morphological distinctions, severe class imbalance (long-tailed distribution), poor image quality in field conditions, and weak cross-domain generalization—this work introduces the first lab-to-field collaborative classification system covering over 230 genera and 1,769 species. We systematically benchmark 12 vision models and propose a novel architecture integrating Vision-Language Transformers with an MLP head, augmented by long-tail learning strategies and a dedicated lab-to-field domain adaptation paradigm. Our method achieves 97% genus-level and 94% species-level accuracy; even with 50% less training data, performance degradation remains under 2%, substantially outperforming baselines. Quantitative analysis confirms significant domain shift between lab and field images, which our approach effectively mitigates. The system delivers a highly sample-efficient and domain-robust solution for automated biodiversity monitoring.

Technology Category

• Machine Learning: Transfer, Domain Adaptation, Multi-Task Learning
• Natural Language Processing: Language Grounding & Multi-modal NLP
• Computer Vision: Multi-modal Vision

Application Category

• Economics, Online Markets and Human Computation: Humans versus LLMs for data annotation and labeling
• Search and Retrieval-Augmented AI: Vertical and domain-specific search
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

Ground beetles are a highly sensitive and speciose biological indicator, making them vital for monitoring biodiversity. However, they are currently an underutilized resource due to the manual effort required by taxonomic experts to perform challenging species differentiations based on subtle morphological differences, precluding widespread applications. In this paper, we evaluate 12 vision models on taxonomic classification across four diverse, long-tailed datasets spanning over 230 genera and 1769 species, with images ranging from controlled laboratory settings to challenging field-collected (in-situ) photographs. We further explore taxonomic classification in two important real-world contexts: sample efficiency and domain adaptation. Our results show that the Vision and Language Transformer combined with an MLP head is the best performing model, with 97% accuracy at genus and 94% at species level. Sample efficiency analysis shows that we can reduce train data requirements by up to 50% with minimal compromise in performance. The domain adaptation experiments reveal significant challenges when transferring models from lab to in-situ images, highlighting a critical domain gap. Overall, our study lays a foundation for large-scale automated taxonomic classification of beetles, and beyond that, advances sample-efficient learning and cross-domain adaptation for diverse long-tailed ecological datasets.

Problem

• Automating taxonomic classification of diverse beetle species using vision models
• Reducing manual effort in species differentiation via machine learning
• Addressing domain adaptation challenges between lab and field images

Innovation

• Vision and Language Transformer with MLP head
• Sample efficiency reduces train data by 50%
• Domain adaptation from lab to in-situ images