Sven Wachsmuth

Qiaoyue Yang, Sven Heutger, Christopher Niemann et al.

Human motion describes the three-dimensional full-body movement of a person. Anticipating such motion holds significant relevance across a wide range of application domains such as human-robot interaction, autonomous driving, animation, and healthcare. In recent research, spatial and temporal dependencies are modeled by bidirectional attention mechanisms. These typically anticipate human motion in an autoregressive manner which could cause an accumulation of errors over time. As a consequence, they solely focus on local pose forecasting. To address these limitations, we propose a non-autoregressive transformer based on spatio-temporal attention, and train it not only for local pose anticipation, but also for global motion prediction in space. Furthermore, to enhance its applicability in real-world scenarios, our model is also trained to recover missing joints due to occlusions, and is capable of processing varying lengths of history observations. Our code is publicly available at https://github.com/Q-Y-Yang/Prediction-of-Local-and-Global-Human-Motion.

Dennis Holzmann, Sven Wachsmuth

We present a novel approach for egocentric action recognition that leverages 2D point tracks as an additional motion cue. While most existing methods rely on RGB appearance, human pose estimation, or their combination, our work demonstrates that tracking randomly sampled image points across video frames can substantially improve recognition accuracy. Unlike prior approaches, we do not detect hands, objects, or interaction regions. Instead, we employ CoTracker to follow a set of randomly initialized points through each video and use the resulting trajectories, together with the corresponding image frames, as input to a Transformer-based recognition model. Surprisingly, our method achieves notable gains even when only the initial frame and its associated point tracks are provided, without incorporating the full video sequence. Experimental results confirm that integrating 2D point tracks consistently enhances performance compared to the same model trained without motion information, highlighting their potential as a lightweight yet effective representation for egocentric action understanding.

Robin Rasch, Sven Wachsmuth, Matthias König

In future, robots will be present in everyday life. The development of these supporting robots is a challenge. A fundamental task for assistance robots is to pick up and hand over objects to humans. By interacting with users, soft factors such as predictability, safety and reliability become important factors for development. Previous works show that collaboration with robots is more acceptable when robots behave and move human-like. In this paper, we present a motion model based on the motion profiles of individual joints. These motion profiles are based on observations and measurements of joint movements in human-human handover. We implemented this joint motion model (JMM) on a humanoid and a non-humanoidal industrial robot to show the movements to subjects. Particular attention was paid to the recognizability and human similarity of the movements. The results show that people are able to recognize human-like movements and perceive the movements of the JMM as more human-like compared to a traditional model. Furthermore, it turns out that the differences between a linear joint space trajectory and JMM are more noticeable in an industrial robot than in a humanoid robot.