Open-Set Cattle Muzzle Identification: A Leakage-Controlled Benchmark and Evaluation Protocol
Not provided in the abstract
Abstract
This paper presents a novel approach to cattle muzzle identification that allows for open-set recognition and incremental enrollment without model retraining.
Reality Card
The study demonstrates that a hybrid CNN-ViT model can achieve high detection-and-identification rates for cattle muzzle identification while allowing for incremental enrollment.
The hybrid model achieves detection-and-identification rates of 98.3%, 96.4%, and 93.6% at target false-acceptance rates of 10^(-1), 10^(-2), and 10^(-3), respectively.
The performance difference between oracle threshold selection and deployable threshold calibration raises concerns about practical applicability.
Paper to code
Verified implementation resources so builders can test the paper’s claims instead of stopping at the abstract.