Meta-Album FNG: 800 Fungi Images from Denmark for Few-Shot Learning

Meta-Album Fungi Dataset (Micro) is a subset of the Danish Fungi 2020 dataset, containing 800 images across 20 classes. The original data was sourced from 3,300 citizen botanists and verified by peers and experts. Images are 128x128 pixels, cropped and scaled from the original repository.

Use Cases

• Benchmarking few-shot learning algorithms based on the standardized 20-class image classification task.
• Training image classifiers for fungi species identification based on the 128x128 pixel images.
• Studying the effects of citizen-science data collection and expert verification on model performance.
• Developing computer vision models for ecological applications within the plant domain.

Strengths

• Includes 800 images, providing a standardized benchmark scale.
• Data quality is supported by a multi-stage verification process involving 3,300 citizen scientists and experts.
• Images are uniformly pre-processed to a 128x128 pixel resolution.

Limitations

• Row count for the full dataset is unknown, limiting suitability assessment for larger-scale projects.
• Column-level documentation is absent; field semantics must be inferred after download.
• The dataset is a significant reduction from the original 128 classes, which may limit taxonomic breadth.

Provenance

Source

Danish Fungi Dataset, sourced from the Atlas of Danish Fungi repository.

Collection Method

Images and labels were collected by citizen botanists, peer-verified via a reliability ranking system, and finally verified by experts.

Freshness

Created on 01 March 2022; last update date is unknown.

Geography

Denmark