Recognising detailed clothing characteristics (fine-grained attributes) in
unconstrained images of people in-the-wild is a challenging task for computer
vision, especially when there is only limited training data from the wild
whilst most data available for model learning are captured