A new visual recognition approach improved a machine learning technique’s ability to both identify an object and how it is oriented in space, according to a study presented in October at the European Conference on Computer Vision in Milan, Italy.
Helping machine learning models identify objects in any pose
