Abstract
Providing systems the ability to relate linguistic and visual content is one of the hallmarks of computer vision. Tasks such as text-based image retrieval and image captioning were designed to test this ability but come with evaluation measures that have a high variance or are difficult to interpret. We study an alternative task for systems that match text and images: given a text query, the system is asked to select the image that best matches the query from a pair of semantically similar images. The system’s accuracy on this Binary Image SelectiON (BISON) task is interpretable, eliminates the reliability problems of retrieval evaluations, and focuses on the system’s ability to understand fine-grained visual structure. We gather a BISON dataset that complements the COCO dataset and use it to evaluate modern text-based image retrieval and image captioning systems. Our results provide novel insights into the performance of these systems. The COCO-BISON dataset and corresponding evaluation code are publicly available from \url{this http URL}.
Abstract (translated by Google)
URL
https://arxiv.org/abs/1901.06595