Theia : An end‐to‐end edge‐ AI mobile framework for high‐throughput field phenotyping and on‐device geometric morphometrics
Methods in Ecology and Evolution · 2026-11 · 17 (11) · 3157-3169
DOI: 10.1111/2041-210x.70416 ↗Kurz gefasst
Fachartikel in der Zeitschrift Methods in Ecology and Evolution, online erschienen am 05.10.2026. Verfasst von Cristobal Bragagnolo und 3 weiteren Autor:innen. Die Arbeit nennt 24 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Ministerio de Ciencia, Innovación y Universidades, Organismo Autónomo de Parques Nacionales.
Abstract (Originalsprache)
Geometric morphometrics (GM) is widely used to study phenotypic variation in ecological and evolutionary research, but the throughput of manual landmark digitization limits how large morphometric datasets can practically become, especially in field studies that would benefit from rapid on‐site feedback. We introduce Theia , an edge‐AI mobile framework that performs end‐to‐end GM workflows directly on‐device, including image acquisition, automated landmark detection and morphometric analysis. The system integrates a dual‐stage deep‐learning pipeline with an offline analytical engine implementing generalized procrustes analysis (GPA) and principal component analysis (PCA), enabling field‐based analyses on low‐cost mobile hardware. Validation against the geomorph R package demonstrated near‐perfect computational fidelity (| r | > 0.999). Comparisons with expert manual digitization across 268 held‐out specimens showed close morphospace concordance, with biological variation accounting for 92.9% of total shape variance and the method effect (AI vs. manual) for 0.5%. The method effect is small, so the shape ordination used in downstream analysis is preserved. Theia enables standardized, reproducible morphometric data collection directly in the field. It complements established desktop workflows and suits studies that need portability and on‐site quality control, such as large‐scale spatial sampling and long‐term monitoring.
Förderung
- Ministerio de Ciencia, Innovación y Universidades
- Organismo Autónomo de Parques Nacionales
Zitieren
APA
Bragagnolo, C., Ferreira‐Rodríguez, A., Abdelaziz, M., & Muñoz‐Pajares, A. J. (2026). Theia : An end‐to‐end edge‐ AI mobile framework for high‐throughput field phenotyping and on‐device geometric morphometrics. Methods in Ecology and Evolution, 17(11), 3157-3169. https://doi.org/10.1111/2041-210x.70416
Vancouver
Bragagnolo C, Ferreira‐Rodríguez A, Abdelaziz M, Muñoz‐Pajares AJ. Theia : An end‐to‐end edge‐ AI mobile framework for high‐throughput field phenotyping and on‐device geometric morphometrics. Methods in Ecology and Evolution. 2026;17(11):3157-3169. doi:10.1111/2041-210x.70416
BibTeX
@article{Bragagnolo2026,
author = {Bragagnolo, Cristobal and Ferreira‐Rodríguez, Andrés and Abdelaziz, Mohamed and Muñoz‐Pajares, A. Jesús},
title = {Theia : An end‐to‐end edge‐ AI mobile framework for high‐throughput field phenotyping and on‐device geometric morphometrics},
journal = {Methods in Ecology and Evolution},
year = {2026},
volume = {17},
number = {11},
pages = {3157--3169},
doi = {10.1111/2041-210x.70416}
}
EUVIMED