A pipeline combining zero-shot detection, motion-aware tracking, and vision transformers achieves 94.2% accuracy on pig behavior recognition tasks, improving 21.2 points over prior methods on the Edinburgh Pig Dataset.
Using a sliding window approach with majority-based filtering, we generated 14,255 temporal windows from the original dataset
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
A Computer Vision Pipeline for Individual-Level Behavior Analysis: Benchmarking on the Edinburgh Pig Dataset
A pipeline combining zero-shot detection, motion-aware tracking, and vision transformers achieves 94.2% accuracy on pig behavior recognition tasks, improving 21.2 points over prior methods on the Edinburgh Pig Dataset.