This study proposes a framework for describing a scene change using natural language text based on indoor scene observations conducted before and after a scene change. The recognition of scene changes plays an essential role in a variety of real-world applications, such as scene anomaly detection. Most scene understanding research has focused on static scenes. Most existing scene change captioning methods detect scene changes from single-view RGB images, neglecting the underlying three-dimensional structures. Previous three-dimensional scene change captioning methods use simulated scenes consisting of geometry primitives, making it unsuitable for real-world applications. To solve these problems, we automatically generated large-scale indoor scene change caption datasets. We propose an end-to-end framework for describing scene changes from various input modalities, namely, RGB images, depth images, and point cloud data, which are available in most robot applications. We conducted experiments with various input modalities and models and evaluated model performance using datasets with various levels of complexity. Experimental results show that the models that combine RGB images and point cloud data as input achieve high performance in sentence generation and caption correctness and are robust for change type understanding for datasets with high complexity. The developed datasets and models contribute to the study of indoor scene change understanding.

This paper proposes a framework that allows the observation of a scene iteratively to answer a given question about the scene. Conventional visual question answering (VQA) methods are designed to answer given questions based on single-view images. However, in real-world applications, such as human–robot interaction (HRI), in which camera angles and occluded scenes must be considered, answering questions based on single-view images might be difficult. Since HRI applications make it possible to observe a scene from multiple viewpoints, it is reasonable to discuss the VQA task in multi-view settings. In addition, because it is usually challenging to observe a scene from arbitrary viewpoints, we designed a framework that allows the observation of a scene actively until the necessary scene information to answer a given question is obtained. The proposed framework achieves comparable performance to a state-of-the-art method in question answering and simultaneously decreases the number of required observation viewpoints by a significant margin. Additionally, we found our framework plausibly learned to choose better viewpoints for answering questions, lowering the required number of camera movements. Moreover, we built a multi-view VQA dataset based on real images. The proposed framework shows high accuracy (94.01%) for the unseen real image dataset.

Summaries of scientific publications enable readers to gain an overview of a large number of studies, but users’ preferences have not yet been explored. In this paper, we conduct two user studies (i.e., short- and long-term studies) where Japanese university students read summaries of English research articles that were either manually written or automatically generated using text summarization and/or machine translation. In the short-term experiment, subjects compared and evaluated the two types of summaries of the same article. We analyze the characteristics in the generated summaries that readers regard as important, such as content richness and simplicity. The experimental results show that subjects are mainly judged based on four criteria, including content richness, simplicity, fluency, and format. In the long-term experiment, subjects read 50 summaries and answered whether they would like to read the original papers after reading the summaries. We discuss the characteristics in the summaries that readers tend to use to determine whether to read the papers, such as topic, methods, and results. The comments from subjects indicate that specific components of scientific publications, including research topics and methods, are important to judge whether to read or not. Our study provides insights to enhance the effectiveness of automatic summarization of scientific publications.