Floralens: a Deep Learning Model for the Portuguese Native Flora

2024-02-13

arXiv.org

AI Summary

High-quality benchmark datasets and lightweight models for native Portuguese plant image recognition are lacking. Method: We introduce Floralens—the first authoritative, multi-source (Botanical Society of Portugal + GBIF), publicly available image dataset for native Portuguese flora—alongside an efficient deep convolutional neural network model trained via Google AutoML Vision. Our methodology includes rigorous data cleaning, cross-source annotation alignment, and an end-to-end open, reproducible pipeline. Contribution/Results: The model achieves accuracy comparable to Pl@ntNet and is deployed online on Project Biolens. Both the dataset and model are openly released via Zenodo. This work fills a critical gap in Iberian Peninsula plant intelligence—providing the first dedicated dataset and deployable model for native flora—and establishes a scalable, citizen-science–enabled paradigm for biodiversity monitoring.

Technology Category

• Machine Learning: Multimodal Learning
• Natural Language Processing: Ethics — Bias, Fairness, Transparency & Privacy
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Web Mining and Content Analysis: Large pretrained models with web data
• Search and Retrieval-Augmented AI: Multilingual and cross-lingual Web search
• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets

Abstract

Machine-learning techniques, especially deep convolutional neural networks, are pivotal for image-based identification of biological species in many Citizen Science platforms. In this paper, we describe the construction of a dataset for the Portuguese native flora based on publicly available research-grade datasets, and the derivation of a high-accuracy model from it using off-the-shelf deep convolutional neural networks. We anchored the dataset in high-quality data provided by Sociedade Portuguesa de Bot^anica and added further sampled data from research-grade datasets available from GBIF. We find that with a careful dataset design, off-the-shelf machine-learning cloud services such as Google's AutoML Vision produce accurate models, with results comparable to those of Pl@ntNet, a state-of-the-art citizen science platform. The best model we derived, dubbed Floralens, has been integrated into the public website of Project Biolens, where we gather models for other taxa as well. The dataset used to train the model is also publicly available on Zenodo.

Problem

• Develops deep learning model for Portuguese flora identification
• Creates dataset from public research-grade botanical sources
• Compares model accuracy with state-of-the-art platforms

Innovation

• Deep convolutional neural networks for species identification
• Dataset construction from public research-grade sources
• Integration with Google's AutoML Vision for accuracy