Meta-Album PLK Mini: 3,440 Plankton Images Across 86 Classes

Meta_Album_PLK_Mini contains 3,440 preprocessed plankton images across 86 classes, sourced from the Woods Hole Oceanographic Institution. The dataset was created by researchers including Heidi M. Sosik, with images captured using an Imaging FlowCytobot and preprocessed to a uniform 128x128 pixel size. The Meta-Album version was authored by Ihsan Ullah and released in March 2022.

Use Cases

• Benchmarking few-shot image classification models based on the 86 plankton classes.
• Training meta-learning algorithms for visual recognition using the standardized 128x128 image format.
• Developing plankton identification tools for ecological monitoring based on the controlled environment images.
• Evaluating image preprocessing techniques for biological specimens based on the described background and blurring method.

Strengths

• Contains 3,440 images, providing a substantial base for few-shot learning tasks.
• Covers 86 distinct plankton classes, offering a multi-class classification challenge.
• Images are preprocessed to a consistent 128x128 pixel size, facilitating model input standardization.
• Has clear provenance from the Woods Hole Oceanographic Institution and a detailed preprocessing methodology.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Last update date is unknown; freshness unverified.
• Data may reflect bias inherent to the specific collection environment and instrument (Imaging FlowCytobot).

Provenance

Source

Woods Hole Oceanographic Institution (WHOI)

Collection Method

Images captured using an Imaging FlowCytobot (IFCB) in a controlled environment, then preprocessed (background addition, Gaussian blur, resizing).