speciesnet-taxonomy-mapping

Map custom camera-trap labels to the SpeciesNet taxonomy by processing a user-provided label list and generating a CSV file that aligns each label with the corresponding SpeciesNet taxonomic hierarchy for precise model predictions.

A SpeciesNet taxonomy mapping CSV is used to filter and adjust MegaDetector model predictions by restricting classification to specific taxa, enabling precise model predictions when working with custom camera-trap category labels.

Use restrict_to_taxa_list by providing a custom camera-trap label list to generate a mapping CSV, which is then passed to MegaDetector to constrain classification results to the specified SpeciesNet taxonomic hierarchy.

Process your own custom camera-trap label list by supplying it to the Skill, which accepts various file types and structures to output a CSV mapping each label to the correct SpeciesNet taxon.

MegaDetector taxonomy mapping supports custom study area or region specifications during the CSV generation process, allowing the mapping to align with the geographical context of your camera-trap data.

Limitations of mapping custom labels to the SpeciesNet taxonomy include the dependency on the existing SpeciesNet taxonomic hierarchy, meaning labels without a direct corresponding taxon may not map correctly for MegaDetector prediction filtering.