A Multimodal Target-Source Classifier With Attention Branches to Understand Ambiguous Instructions for Fetching Daily Objects.
Aly MagassoubaKomei SugiuraHisashi KawaiPublished in: IEEE Robotics Autom. Lett. (2020)
Keyphrases
- focus of attention
- target object
- feature values
- target images
- training data
- feature space
- multiple objects
- d objects
- data objects
- multi modal
- feature selection
- previously learned
- neural network
- training set
- support vector
- learning algorithm
- semantic labels
- classification process
- linear classifiers
- classification scheme
- visual attention
- vision system
- image features