Prototypicality effects in global semantic description of objects
This addresses the problem of interpretable semantic description in computer vision, though it appears incremental as it builds on existing CNN and prototype theory concepts.
The paper tackles semantic description of object features by introducing a prototype-based model that encodes semantic meaning using CNN-classification models, resulting in discriminative descriptor signatures that preserve semantic information, enable typicality scoring, and simulate prototypical organization within categories.
In this paper, we introduce a novel approach for semantic description of object features based on the prototypicality effects of the Prototype Theory. Our prototype-based description model encodes and stores the semantic meaning of an object, while describing its features using the semantic prototype computed by CNN-classifications models. Our method uses semantic prototypes to create discriminative descriptor signatures that describe an object highlighting its most distinctive features within the category. Our experiments show that: i) our descriptor preserves the semantic information used by the CNN-models in classification tasks; ii) our distance metric can be used as the object's typicality score; iii) our descriptor signatures are semantically interpretable and enables the simulation of the prototypical organization of objects within a category.