MobileMold: 4,941 Smartphone Microscopy Images for Food Mold Detection

4,941 annotated images for food mold detection, captured using smartphones with clip-on microscope attachments. The dataset was created by Dinh Nam Pham and colleagues and was presented at the ACM Multimedia Systems Conference in April 2026. It supports research into accessible, low-cost food safety monitoring via computer vision.

Use Cases

• Train image classification models for mold detection based on annotated smartphone microscopy images.
• Develop mobile applications for on-site food quality inspection based on smartphone-based microscopy.
• Benchmark model performance for mold detection on various food surfaces as described in the dataset scope.

Strengths

• Contains 4,941 annotated images, providing a substantial base for model training.
• Images are captured using accessible smartphone-based microscopy, addressing a need for low-cost monitoring.
• Has an associated peer-reviewed publication (MMSys '26) and a dedicated website for documentation.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Data may reflect bias inherent to the specific smartphone and microscopy attachment setup used.

Provenance

Source

figshare, with data available via Zenodo DOI.

Collection Method

Images captured using smartphones with various clip-on microscope attachments.

Freshness

Last updated 2026-04-28 11:32:16; freshness should be verified.