Datasets:

imageomics

/

plazi-biospecimen-img-descriptions

like 2

Follow

HDR Imageomics Institute 65

Table of figures and their captions. The table in TSV format was generated from the Blazegraph journal included here, and has 292096 rows. The table in CSV format was obtained from Plazi's public SPARQL endpoint (using curl), and has 298260 rows.

Both tables were generated using the same SPARQL query, and are non-redundant for figures and their figure image URLs.

The columns of the tables are as follows, as documented in the SPARQL query:

• MD5 hash of figure URI as filename-friendly ID that is easy to back-reference to this table
• figure URI (note that in the TSV table this is enclosed in angle brackets)
• URL of the figure's image representation
• textual caption of the figure

To achieve non-redundancy, the extraction query considers only figures that have only one figure representation, and only one caption. In addition, the query disregards figures with an empty caption, and figures that don't seem to be locatable online (as inferred from having a file:/// URI).