Learn to Predict How Humans Manipulate Large-sized Objects from Interactive Motions
This work addresses a gap in understanding human-object interactions for applications like human-robot collaboration, but it is incremental as it builds on existing prediction methods with a new dataset and descriptors.
The paper tackles the problem of predicting future states of humans and large-sized daily objects during full-body interactions, achieving state-of-the-art results by incorporating object dynamic descriptors into a graph neural network, which improves generalization to unseen objects.
Understanding human intentions during interactions has been a long-lasting theme, that has applications in human-robot interaction, virtual reality and surveillance. In this study, we focus on full-body human interactions with large-sized daily objects and aim to predict the future states of objects and humans given a sequential observation of human-object interaction. As there is no such dataset dedicated to full-body human interactions with large-sized daily objects, we collected a large-scale dataset containing thousands of interactions for training and evaluation purposes. We also observe that an object's intrinsic physical properties are useful for the object motion prediction, and thus design a set of object dynamic descriptors to encode such intrinsic properties. We treat the object dynamic descriptors as a new modality and propose a graph neural network, HO-GCN, to fuse motion data and dynamic descriptors for the prediction task. We show the proposed network that consumes dynamic descriptors can achieve state-of-the-art prediction results and help the network better generalize to unseen objects. We also demonstrate the predicted results are useful for human-robot collaborations.