Meta-Album FNG: 25 Fungus Species Images from Denmark

15,122 images of fungi from 25 classes, sampled from the Danish Fungi 2020 dataset. The data was sourced and verified by a network of 3,300 citizen botanists and experts from the Atlas of Danish Fungi. Each image is a 128x128 color crop of a fungus, standardized from the original dataset.

Use Cases

• Few-shot image classification based on 25 distinct fungus species.
• Training computer vision models for automated plant and fungi recognition.
• Benchmarking model performance on a dataset with expert-verified citizen-science labels.
• Studying image preprocessing techniques like center-cropping and resizing for biological specimens.

Strengths

• Contains 15,122 images with a defined class distribution (minimum 372, maximum 1,221 images per class).
• Data quality is enhanced by a multi-stage verification process involving 3,300 citizen scientists and expert review.
• Images are uniformly preprocessed to a standard 128x128 pixel size using an anti-aliasing filter.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• The dataset is a curated subset of 25 classes from a larger source, which may limit generalizability.
• Images were cropped to a square from varying original sizes, which may have removed contextual information.

Provenance

Source

Danish Fungi 2020 dataset, sourced from the Atlas of Danish Fungi repository.

Collection Method

Images and labels were collected by 3,300 citizen botanists, peer-verified via a reliability ranking system, and finally verified by experts.

Freshness

Created 01 March 2022; last update date is unknown.

Geography

Denmark