Abstract We present ZeroEGGS, a neural network framework for speech‐driven gesture generation with zero‐shot style control by example. This means style can be controlled via only a short example motion clip, even for motion styles unseen during training. Our model uses a Variational framework to learn a style embedding, making it easy to modify style through latent space manipulation or blending and scaling of style embeddings. The probabilistic nature of our framework further enables the generation of a variety of outputs given the input, addressing the stochastic nature of gesture motion. In a series of experiments, we first demonstrate the flexibility and generalizability of our model to new speakers and styles. In a user study, we then show that our model outperforms previous state‐of‐the‐art techniques in naturalness of motion, appropriateness for speech, and style portrayal. Finally, we release a high‐quality dataset of full‐body gesture motion including fingers, with speech, spanning across 19 different styles. Our code and data are publicly available at https://github.com/ubisoft/ubisoft‐laforge‐ZeroEGGS .

Exposure to nature has been shown to have a positive effect on people’s mental health. Little research has compared restorative effects of simulated nature presented by different media. Here, we investigated stress recovery when viewing a computer-generated nature setting presented in visual and pictorial space in virtual reality. Participants experienced a stress induction task and were then put into one of two relaxation scenarios: they either viewed the nature scene in visual space, (they were immersed into it; presence condition), or they viewed a large depiction of it in pictorial space (picture condition). Participants’ affective state was assessed before and after stress induction, and after relaxation using the ZIPERS questionnaire. We additionally recorded electrodermal activity as a measure of physiological arousal. The results revealed that relaxation led to an increase in positive affect scores and a decrease in electrodermal activity only in the presence condition. The negative affect scores decreased significantly for both conditions similarly. Our results show that restoration is more effective in visual than in pictorial space.

Gait recognition, referring to the identification of individuals based on the manner in which they walk, can be very challenging due to the variations in the viewpoint of the camera and the appearance of individuals. Current methods for gait recognition have been dominated by deep learning models, notably those based on partial feature representations. In this context, we propose a novel deep network, learning to transfer multi-scale partial gait representations using capsules to obtain more discriminative gait features. Our network first obtains multi-scale partial representations using a state-of-the-art deep partial feature extractor. It then recurrently learns the correlations and co-occurrences of the patterns among the partial features in forward and backward directions using Bidirectional Gated Recurrent Units (BGRU). Finally, a capsule network is adopted to learn deeper part-whole relationships and assigns more weights to the more relevant features while ignoring the spurious dimensions. That way, we obtain final features that are more robust to both viewing and appearance changes. The performance of our method has been extensively tested on two gait recognition datasets, CASIA-B and OU-MVLP, using four challenging test protocols. The results of our method have been compared to the state-of-the-art gait recognition solutions, showing the superiority of our model, notably when facing challenging viewing and carrying conditions.

Abstract We present a probabilistic framework to generate character animations based on weak control signals, such that the synthesized motions are realistic while retaining the stochastic nature of human movement. The proposed architecture, which is designed as a hierarchical recurrent model, maps each sub‐sequence of motions into a stochastic latent code using a variational autoencoder extended over the temporal domain. We also propose an objective function which respects the impact of each joint on the pose and compares the joint angles based on angular distance. We use two novel quantitative protocols and human qualitative assessment to demonstrate the ability of our model to generate convincing and diverse periodic and non‐periodic motion sequences without the need for strong control signals.

Large datasets are the cornerstone of recent advances in computer vision using deep learning. In contrast, existing human motion capture (mocap) datasets are small and the motions limited, hampering progress on learning models of human motion. While there are many different datasets available, they each use a different parameterization of the body, making it difficult to integrate them into a single meta dataset. To address this, we introduce AMASS, a large and varied database of human motion that unifies 15 different optical marker-based mocap datasets by representing them within a common framework and parameterization. We achieve this using a new method, MoSh++, that converts mocap data into realistic 3D human meshes represented by a rigged body model. Here we use SMPL [Loper et al., 2015], which is widely used and provides a standard skeletal representation as well as a fully rigged surface mesh. The method works for arbitrary marker sets, while recovering soft-tissue dynamics and realistic hand motion. We evaluate MoSh++ and tune its hyperparameters using a new dataset of 4D body scans that are jointly recorded with marker-based mocap. The consistent representation of AMASS makes it readily useful for animation, visualization, and generating training data for deep learning. Our dataset is significantly richer than previous human motion collections, having more than 40 hours of motion data, spanning over 300 subjects, more than 11000 motions, and will be publicly available to the research community.

The individual shape of the human body, including the geometry of its articulated structure and the distribution of weight over that structure, influences the kinematics of a person’s movements. How sensitive is the visual system to inconsistencies between shape and motion introduced by retargeting motion from one person onto the shape of another? We used optical motion capture to record five pairs of male performers with large differences in body weight, while they pushed, lifted, and threw objects. From these data, we estimated both the kinematics of the actions as well as the performer’s individual body shape. To obtain consistent and inconsistent stimuli, we created animated avatars by combining the shape and motion estimates from either a single performer or from different performers. Using these stimuli we conducted three experiments in an immersive virtual reality environment. First, a group of participants detected which of two stimuli was inconsistent. Performance was very low, and results were only marginally significant. Next, a second group of participants rated perceived attractiveness, eeriness, and humanness of consistent and inconsistent stimuli, but these judgements of animation characteristics were not affected by consistency of the stimuli. Finally, a third group of participants rated properties of the objects rather than of the performers. Here, we found strong influences of shape-motion inconsistency on perceived weight and thrown distance of objects. This suggests that the visual system relies on its knowledge of shape and motion and that these components are assimilated into an altered perception of the action outcome. We propose that the visual system attempts to resist inconsistent interpretations of human animations. Actions involving object manipulations present an opportunity for the visual system to reinterpret the introduced inconsistencies as a change in the dynamics of an object rather than as an unexpected combination of body shape and body motion.

Avatars are important for games and immersive social media applications. Although avatars are still not complete digital copies of the user, they often aim to represent a user in terms of appearance (color and shape) and motion. Previous studies have shown that humans can recognize their own motions in point-light displays. Here, we investigated whether recognition of self-motion is dependent on the avatar's fidelity and the congruency of the avatar's sex with that of the participants. Participants performed different actions that were captured and subsequently remapped onto three different body representations: a point-light figure, a male, and a female virtual avatar. In the experiment, participants viewed the motions displayed on the three body representations and responded to whether the motion was their own. Our results show that there was no influence of body representation on self-motion recognition performance, participants were equally sensitive to recognize their own motion on the point-light figure and the virtual characters. In line with previous research, recognition performance was dependent on the action. Sensitivity was highest for uncommon actions, such as dancing and playing ping-pong, and was around chance level for running, suggesting that the degree of individuality of performing certain actions affects self-motion recognition performance. Our results show that people were able to recognize their own motions even when individual body shape cues were completely eliminated and when the avatar's sex differed from own. This suggests that people might rely more on kinematic information rather than shape and sex cues for recognizing own motion. This finding has important implications for avatar design in game and immersive social media applications.

The individual shape of the human body, including the geometry of its articulated structure and the distribution of weight over that structure, influences the kinematics of a person's movements. How sensitive is the visual system to inconsistencies between shape and motion introduced by retargeting motion from one person onto the shape of another? We used optical motion capture to record five pairs of male performers with large differences in body weight, while they pushed, lifted, and threw objects. Based on a set of 67 markers, we estimated both the kinematics of the actions as well as the performer's individual body shape. To obtain consistent and inconsistent stimuli, we created animated avatars by combining the shape and motion estimates from either a single performer or from different performers. In a virtual reality environment, observers rated the perceived weight or thrown distance of the objects. They were also asked to explicitly discriminate between consistent and hybrid stimuli. Observers were unable to accomplish the latter, but hybridization of shape and motion influenced their judgements of action outcome in systematic ways. Inconsistencies between shape and motion were assimilated into an altered perception of the action outcome.

We present and evaluate a method of reconstructing three-dimensional (3D) periodic human motion from two-dimensional (2D) motion sequences. Using Fourier decomposition, we construct a compact representation for periodic human motion. A low-dimensional linear motion model is learned from a training set of 3D Fourier representations by means of principal components analysis. Two-dimensional test data are projected onto this model with two approaches: least-square minimization and calculation of a maximum a posteriori probability using the Bayes' rule. We present two different experiments in which both approaches are applied to 2D data obtained from 3D walking sequences projected onto a plane. In the first experiment, we assume the viewpoint is known. In the second experiment, the horizontal viewpoint is unknown and is recovered from the 2D motion data. The results demonstrate that by using the linear model, not only can missing motion data be reconstructed, but unknown view angles for 2D test data can also be retrieved.

We present empirical studies that consider the effects of stereopsis and simulated aerial perspective on depth perception in translucent volumes. We consider a purely absorptive lighting model, in which light is not scattered or reflected, but is simply absorbed as it passes through the volume. A purely absorptive lighting model is used, for example, when rendering digitally reconstructed radiographs (DRRs), which are synthetic X-ray images reconstructed from CT volumes. Surgeons make use of DRRs in planning and performing operations, so an improvement of depth perception in DRRs may help diagnosis and surgical planning.

In this paper, we present Paper Windows, a prototype windowing environment that simulates the use of digital paper displays. By projecting windows on physical paper, Paper Windows allows the capturing of physical affordances of paper in a digital world. The system uses paper as an input device by tracking its motion and shape with a Vicon Motion Capturing System. We discuss the design of a number of interaction techniques for manipulating information on paper displays.