| dc.description.abstract | Body condition score is one of the important parameters in dairy practice as it can efficiently
estimate a cow’s energy balance. However, due to the limitations of the current manual scoring
system, there is growing interest in the automatization of the condition scoring system.
Therefore, our study is aimed for estimate the accuracy of a convolutional neural network
trained by supervised machine learning that could recall the condition estimated by
professionals. The images were recorded from 3 large-scale dairy farms using a simple 2D
camera that faces the cow’s rump. The images were annotated by the same professional with a
bounding box with the classification of 12 classes. | en_US |
