Optimal deep learning pipeline for animal activity recognition using bio-logging data

Summary

This article presents a comprehensive scientific study on developing an optimal deep learning pipeline for animal activity recognition using bio-logging data (accelerometers, gyroscopes, magnetometers, GPS). It covers the entire workflow from data collection via animal-borne devices, through preprocessing and feature engineering, to training and evaluating supervised machine learning models for behaviour classification in wildlife and livestock. The paper likely benchmarks various deep learning architectures (CNNs, RNNs, transformers) and provides best-practice recommendations for researchers in the field of animal behaviour ecology and precision livestock farming.

Source

bskyOptimal deep learning pipeline for animal activity recognition using bio-logging databesjournals.onlinelibrary.wiley.com