An Ecologist‐Friendly R Workflow for Expediting Species‐Level Classification of Camera Trap Images

L. Petroni, L. Natucci, A. Massolo · 2024

Read An Ecologist‐Friendly R Workflow for Expediting Species‐Level Classification of Camera Trap Images by L. Petroni, L. Natucci, A. Massolo (2024) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/pmc-11638126.txt.

Description

L Petroni, L Natucci, and A Massolo offer ecologists an R-based workflow for sorting camera trap photos by species, in a 2024 paper. Camera traps gather huge image sets that are slow to review by hand. Deep learning can help, but it usually demands Python skills. Their workflow keeps the whole job inside R, a language ecologists know well.

An object detector first flags images containing animals. A modern detection model is then trained for the user's own species. Offline image tricks can balance rare and common classes. Results land in a database that pairs with review software for quality checks. A test on mammals of central Italy used only a thousand training photos per species. Scores were high, and most new images were sorted correctly. The scripts and starter models are shared for reuse.

Catalog record

CREATORL. Petroni, L. Natucci, A. Massolo
YEAR2024
COLLECTIONPAPERS
FORMATTXT
FILE SIZE24 KB
VERSION1.0
LICENSECC BY 4.0
RIGHTSCC BY
SOURCEEurope PMC

Source: https://europepmc.org/articles/PMC11638126

Browse the full library · About the Tower of Babel