Open-Source AI Ensemble Revolutionizes Wildlife Monitoring with Efficient Species Detection

An open-source AI ensemble combining object detection and species classification models, trained on over 65M images, revolutionizes wildlife monitoring by efficiently detecting and classifying species in camera trap images, minimizing human review effort.

• The repository hosts code for running an AI ensemble that detects and classifies wildlife species in camera trap images
• The ensemble combines an object detector (MegaDetector) and a species classifier (SpeciesNet) trained on over 65M images
• The ensemble uses heuristics and optional geographic information to assign each image a single category, optimized for minimizing human review effort