Learning Visually Consistent Label Embeddings for Zero-Shot Learning
This addresses the problem of recognizing unseen classes in computer vision, though it appears incremental as it builds on existing zero-shot learning methods.
The paper tackles zero-shot learning by jointly learning visually consistent word vectors and a label embedding model to improve knowledge transfer between classes, achieving significant improvements in recognition accuracy on two benchmark datasets.
In this work, we propose a zero-shot learning method to effectively model knowledge transfer between classes via jointly learning visually consistent word vectors and label embedding model in an end-to-end manner. The main idea is to project the vector space word vectors of attributes and classes into the visual space such that word representations of semantically related classes become more closer, and use the projected vectors in the proposed embedding model to identify unseen classes. We evaluate the proposed approach on two benchmark datasets and the experimental results show that our method yields significant improvements in recognition accuracy.