Visual Question Answering on Image Sets
This work addresses the problem of visual question answering across multiple images for researchers in computer vision, but it is incremental as it generalizes an existing task.
The authors introduced Image-Set Visual Question Answering (ISVQA), a task extending single-image VQA to multi-image settings, and created two datasets with 141,096 total questions to enable research, analyzing their properties and building baseline models.
We introduce the task of Image-Set Visual Question Answering (ISVQA), which generalizes the commonly studied single-image VQA problem to multi-image settings. Taking a natural language question and a set of images as input, it aims to answer the question based on the content of the images. The questions can be about objects and relationships in one or more images or about the entire scene depicted by the image set. To enable research in this new topic, we introduce two ISVQA datasets - indoor and outdoor scenes. They simulate the real-world scenarios of indoor image collections and multiple car-mounted cameras, respectively. The indoor-scene dataset contains 91,479 human annotated questions for 48,138 image sets, and the outdoor-scene dataset has 49,617 questions for 12,746 image sets. We analyze the properties of the two datasets, including question-and-answer distributions, types of questions, biases in dataset, and question-image dependencies. We also build new baseline models to investigate new research challenges in ISVQA.