Resolving References to Objects in Photographs using the Words-As-Classifiers Model
This work addresses the challenge of linking language to visual perception for computers, with potential applications in image understanding and human-computer interaction, though it is incremental as it extends an existing model to new datasets.
The paper tackled the problem of resolving references to objects in photographs using a grounded semantics model, achieving performance competitive with state-of-the-art methods in reference resolution tasks.
A common use of language is to refer to visually present objects. Modelling it in computers requires modelling the link between language and perception. The "words as classifiers" model of grounded semantics views words as classifiers of perceptual contexts, and composes the meaning of a phrase through composition of the denotations of its component words. It was recently shown to perform well in a game-playing scenario with a small number of object types. We apply it to two large sets of real-world photographs that contain a much larger variety of types and for which referring expressions are available. Using a pre-trained convolutional neural network to extract image features, and augmenting these with in-picture positional information, we show that the model achieves performance competitive with the state of the art in a reference resolution task (given expression, find bounding box of its referent), while, as we argue, being conceptually simpler and more flexible.