A Unified Semantic Embedding: Relating Taxonomies and Attributes
This work addresses the challenge of integrating multiple semantic entities for better categorization in computer vision, though it appears incremental by building on prior methods that used such entities as side information.
The paper tackles the problem of object categorization by learning a unified semantic embedding space that includes categories, supercategories, and attributes, enabling category representation as a linear combination of these entities with improved discriminative composition.
We propose a method that learns a discriminative yet semantic space for object categorization, where we also embed auxiliary semantic entities such as supercategories and attributes. Contrary to prior work which only utilized them as side information, we explicitly embed the semantic entities into the same space where we embed categories, which enables us to represent a category as their linear combination. By exploiting such a unified model for semantics, we enforce each category to be represented by a supercategory + sparse combination of attributes, with an additional exclusive regularization to learn discriminative composition.