Parallel Attention: A Unified Framework for Visual Object Discovery through Dialogs and Queries

TL;DR

The paper introduces the PLAN network, a unified, explainable framework that uses dual recurrent attention mechanisms to improve object discovery in images based on natural language expressions, including dialogs.

Contribution

It proposes a novel parallel attention framework that jointly relates language and visual content for referring expression comprehension, outperforming existing methods.

Findings

Outperforms state-of-the-art on RefCOCO, RefCOCO+, and GuessWhat?! datasets.

Handles variable-length language inputs, from short phrases to multi-round dialogs.

Provides visualizable and explainable attention mechanisms.

Abstract

Recognising objects according to a pre-defined fixed set of class labels has been well studied in the Computer Vision. There are a great many practical applications where the subjects that may be of interest are not known beforehand, or so easily delineated, however. In many of these cases natural language dialog is a natural way to specify the subject of interest, and the task achieving this capability (a.k.a, Referring Expression Comprehension) has recently attracted attention. To this end we propose a unified framework, the ParalleL AttentioN (PLAN) network, to discover the object in an image that is being referred to in variable length natural expression descriptions, from short phrases query to long multi-round dialogs. The PLAN network has two attention mechanisms that relate parts of the expressions to both the global visual content and also directly to object candidates. Furthermore, the attention mechanisms are recurrent, making the referring process visualizable and explainable. The attended information from these dual sources are combined to reason about the referred object. These two attention mechanisms can be trained in parallel and we find the combined system outperforms the state-of-art on several benchmarked datasets with different length language input, such as RefCOCO, RefCOCO+ and GuessWhat?!.