Gesture Generation Research Articles

Personality Expression using Co-speech Gesture

We express our personality through verbal and nonverbal behavior. While verbal cues are mostly related to the semantics of what we say, nonverbal cues include our posture, gestures, and facial expressions. Appropriate expression of these behavioral elements improves conversational virtual agents’ communication capabilities and realism. Although previous studies focus on co-speech gesture generation, they do not consider the personality aspect of the synthesized animations. We show that automatically generated co-speech gestures naturally express personality traits, and heuristics-based adjustments for such animations can further improve personality expression. To this end, we present a framework for enhancing co-speech gestures with the different personalities of the Five-Factor model. Our experiments suggest that users perceive increased realism and improved personality expression when combining heuristics-based motion adjustments with co-speech gestures.

TAG2G: A Diffusion-Based Approach to Interlocutor-Aware Co-Speech Gesture Generation

Extended reality (XR) systems are about to be integrated into our daily lives and will provide support in a variety of fields such as education and coaching. Enhancing user experience demands agents that are capable of displaying realistic affective and social behaviors within these systems, and, as a prerequisite, with the capability of understanding their interaction partner and responding appropriately. Based on our literature review of recent works published in the field of co-speech gesture generation, researchers have developed complex models capable of generating gestures characterized by a high level of human-likeness and speaker appropriateness. Nevertheless, this is only true in settings where the agent has an active status (i.e., the agent acts as the speaker), or it is delivering a monologue in a non-interactive setting. However, as illustrated in multiple works and competitions like the GENEA Challenge, these models remain inadequate in generating interlocutor-aware gestures. We consider interlocutor-aware gesture generation the process of displaying gestures that take into account the conversation partner’s behavior. Moreover, in settings where the agent is the listener, generated gestures lack the level of naturalness that we expect from a face-to-face conversation. To overcome these issues, we have designed a pipeline, called TAG2G, composed of a diffusion model, which was demonstrated to be a stable and powerful tool in gesture generation, and a vector-quantized variational auto-encoder (VQVAE), widely employed to produce meaningful gesture embeddings. Refocusing from monadic to dyadic multimodal input settings (i.e., taking into account text, audio, and previous gestures of both participants of a conversation) allows us to explore and infer the complex interaction mechanisms that lie in a balanced two-sided conversation. As per our results, a multi-agent conversational input setup improves the generated gestures’ appropriateness with respect to the conversational counterparts. Conversely, when the agent is speaking, a monadic approach performs better in terms of the generated gestures’ appropriateness in relation to the speech.

Editable Co-Speech Gesture Synthesis Enhanced with Individual Representative Gestures

Co-speech gesture synthesis is a challenging task due to the complexity and uncertainty between gestures and speech. Gestures that accompany speech (i.e., Co-Speech Gesture) are an essential part of natural and efficient embodied human communication, as they work in tandem with speech to convey information more effectively. Although data-driven approaches have improved gesture synthesis, existing deep learning-based methods use deterministic modeling which could lead to averaging out predicted gestures. Additionally, these methods lack control over gesture generation such as user editing of generated results. In this paper, we propose an editable gesture synthesis method based on a learned pose script, which disentangles gestures into individual representative and rhythmic gestures to produce high-quality, diverse and realistic poses. Specifically, we first detect the time of occurrence of gestures in video sequences and transform them into pose scripts. Regression models are then built to predict the pose scripts. Next, learned pose scripts are used for gesture synthesis, while rhythmic gestures are modeled using a variational auto-encoder and a one-dimensional convolutional network. Moreover, we introduce a large-scale Chinese co-speech gesture synthesis dataset with multimodal annotations for training and evaluation, which will be publicly available to facilitate future research. The proposed method allows for the re-editing of generated results by changing the pose scripts for applications such as interactive digital humans. The experimental results show that this method generates more quality, more diverse, and realistic gestures than other existing methods.

Audio2Gestures: Generating Diverse Gestures From Audio.

People may perform diverse gestures affected by various mental and physical factors when speaking the same sentences. This inherent one-to-many relationship makes co-speech gesture generation from audio particularly challenging. Conventional CNNs/RNNs assume one-to-one mapping, and thus tend to predict the average of all possible target motions, easily resulting in plain/boring motions during inference. So we propose to explicitly model the one-to-many audio-to-motion mapping by splitting the cross-modal latent code into shared code and motion-specific code. The shared code is expected to be responsible for the motion component that is more correlated to the audio while the motion-specific code is expected to capture diverse motion information that is more independent of the audio. However, splitting the latent code into two parts poses extra training difficulties. Several crucial training losses/strategies, including relaxed motion loss, bicycle constraint, and diversity loss, are designed to better train the VAE. Experiments on both 3D and 2D motion datasets verify that our method generates more realistic and diverse motions than previous state-of-the-art methods, quantitatively and qualitatively. Besides, our formulation is compatible with discrete cosine transformation (DCT) modeling and other popular backbones (i.e., RNN, Transformer). As for motion losses and quantitative motion evaluation, we find structured losses/metrics (e.g. STFT) that consider temporal and/or spatial context complement the most commonly used point-wise losses (e.g. PCK), resulting in better motion dynamics and more nuanced motion details. Finally, we demonstrate that our method can be readily used to generate motion sequences with user-specified motion clips on the timeline.

Evaluating Gesture Generation in a Large-scale Open Challenge: The GENEA Challenge 2022

This article reports on the second GENEA Challenge to benchmark data-driven automatic co-speech gesture generation. Participating teams used the same speech and motion dataset to build gesture-generation systems. Motion generated by all these systems was rendered to video using a standardised visualisation pipeline and evaluated in several large, crowdsourced user studies. Unlike when comparing different research articles, differences in results are here only due to differences between methods, enabling direct comparison between systems. The dataset was based on 18 hours of full-body motion capture, including fingers, of different persons engaging in a dyadic conversation. Ten teams participated in the challenge across two tiers: full-body and upper-body gesticulation. For each tier, we evaluated both the human-likeness of the gesture motion and its appropriateness for the specific speech signal. Our evaluations decouple human-likeness from gesture appropriateness, which has been a difficult problem in the field. The evaluation results show some synthetic gesture conditions being rated as significantly more human-like than 3D human motion capture. To the best of our knowledge, this has not been demonstrated before. On the other hand, all synthetic motion is found to be vastly less appropriate for the speech than the original motion-capture recordings. We also find that conventional objective metrics do not correlate well with subjective human-likeness ratings in this large evaluation. The one exception is the Fréchet gesture distance (FGD), which achieves a Kendall’s tau rank correlation of around -0.5. Based on the challenge results we formulate numerous recommendations for system building and evaluation.

Priming users with babies’ gestures: Investigating the influences of priming with different development origin of image schemas in gesture elicitation study

Gesture elicitation study is an effective method to design gestures for various contexts. Through involving end-users, GES results in intuitive gestures because they directly reflect end-users’ mental models and preferences. However, limited by personal experience, end-users are not capable of taking full advantages of technology while proposing gestures, which is referred as legacy bias. To overcome this, previous studies demonstrate that users’ performance can be improved by priming, such as viewing gestures, watching fictional movies, and experiencing framed scenarios. This research extends this line of studies by considering the developmental origin of image schemas in priming. More specifically, we compared the influences of no-priming, priming with early image schemas (EIS), and priming with late image schemas (LIS) on GES. Controlled experiments were conducted (N = 120) along the three stages of GES: users’ generation of gestures (Experiment 1), final gesture sets (Experiment 2), and end-users’ learnability of gestures (Experiment 3). Results show that users are largely influenced by developmental origin of image schemas in priming. LIS-priming improve gesture proposal production in comparison to no-priming condition. As for end-users’ evaluation, EIS-priming gestures exhibit higher initial and overall learnability.

Learning hierarchical discrete prior for co-speech gesture generation

In the context of Co-Speech Gesture Generation, Vector-Quantized Variational Autoencoder (VQ-VAE) based methods have shown promising results by separating the generation process into two stages: learning discrete gesture priors via pretraining for gesture reconstruction, which encodes gesture into a discrete codebook, followed by learning the mapping between speech audio and gesture codebook indices. This design leverages pretraining of motion VQVAE with the motion reconstruction task to improve the quality of generated gestures. However, the vanilla VQVAE’s codebook often fails to encode both low-level and high-level gesture features adequately, resulting in limited reconstruction quality and generation performance. To address this, we propose the Hierarchical Discrete Audio-to-Gesture (HD-A2G), which innovates (i) a two-stage hierarchical codebook structure for capturing high-level and low-level gesture priors, enabling the reconstruction of gesture details. (ii) it further integrates high-level and low-level feature using an AdaIn layer, effectively enhancing the learning of gesture’s rhythm and content. (iii) it explicitly maps text and audio onset features to the appropriate levels of the codebook, ensuring learning accurate hierarchical associations for the generation stage. Experimental results on the BEAT and Trinity datasets demonstrate that HD-A2G outperform the baseline method in both pretrained gesture reconstruction and audio-conditioned gesture generation with a clear margin, achieving the state-of-the-art performance qualitatively and quantitatively.

DiT-Gesture: A Speech-Only Approach to Stylized Gesture Generation

The generation of co-speech gestures for digital humans is an emerging area in the field of virtual human creation. Prior research has progressed by using acoustic and semantic information as input and adopting a classification method to identify the person’s ID and emotion for driving co-speech gesture generation. However, this endeavor still faces significant challenges. These challenges go beyond the intricate interplay among co-speech gestures, speech acoustic, and semantics; they also encompass the complexities associated with personality, emotion, and other obscure but important factors. This paper introduces “DiT-Gestures”, a speech-conditional diffusion-based and non-autoregressive transformer-based generative model with the WavLM pre-trained model and a dynamic mask attention network (DMAN). It can produce individual and stylized full-body co-speech gestures by only using raw speech audio, eliminating the need for complex multimodal processing and manual annotation. Firstly, considering that speech audio contains acoustic and semantic features and conveys personality traits, emotions, and more subtle information related to accompanying gestures, we pioneer the adaptation of WavLM, a large-scale pre-trained model, to extract the style from raw audio information. Secondly, we replace the causal mask by introducing a learnable dynamic mask for better local modeling in the neighborhood of the target frames. Extensive subjective evaluation experiments are conducted on the Trinity, ZEGGS, and BEAT datasets to confirm WavLM’s and the model’s ability to synthesize natural co-speech gestures with various styles.

Improving diversity of speech‐driven gesture generation with memory networks as dynamic dictionaries

AbstractGenerating co‐speech gestures for interactive digital humans remains challenging because of the indeterministic nature of the problem. The authors observe that gestures generated from speech audio or text by existing neural methods often contain less movement shift than expected, which can be viewed as slow or dull. Thus, a new generative model coupled with memory networks as dynamic dictionaries for speech‐driven gesture generation with improved diversity is proposed. More specifically, the dictionary network dynamically stores connections between text and pose features in a list of key‐value pairs as the memory for the pose generation network to look up; the pose generation network then merges the matching pose features and input audio features for generating the final pose sequences. To make the improvements more accurately measurable, a new objective evaluation metric for gesture diversity that can remove the influence of low‐quality motions is also proposed and tested. Quantitative and qualitative experiments demonstrate that the proposed architecture succeeds in generating gestures with improved diversity.

Gesture generation by the robotic hand for aiding speech and hard of hearing persons based on indian sign language

Speech and hearing impairments are among the most common problems in Indian societies. It can affect anyone, whether children, adults, or more. Many different treatments can help to overcome hearing problems. Different types of hearing aids and cochlear implants help amplify sounds for better hearing. The type of language known as sign language is very scientific and has its grammar and syntax. Still, due to a need for more awareness among hard-of-hearing persons, they need to be made familiar with the institutions where they can learn and equip themselves for communication. This paper describes an approach to aid speech and hard-of-hearing persons so that they are free to communicate with persons who do not have speech and hearing disabilities based on the Indian Sign Language System. To find an appropriate solution, there is a need to develop a system that can act as an interpreter for speech and hard-of-hearing persons. The interpreter system is designed with the help of the Robotic hand model and is programmed using Raspberry Pi 4. Based on the experimental results, it can be observed that the robotic hands generated different signs of the alphabet corresponding to the speech commands uttered by an individual. Several experimental trials were conducted by ten persons who do not have any hearing disabilities. The results of the five experimental trials are shown in this paper. The estimation of performance parameters and statistical analysis are also carried out to analyze better and interpret the experimental results. Based on the experimental results, the proposed robotic hand interpreter system model accurately generates gestures corresponding to different alphabets used in the Indian Sign Language system, yielding an overall accuracy of 94 percent.

Chain of Generation: Multi-Modal Gesture Synthesis via Cascaded Conditional Control

This study aims to improve the generation of 3D gestures by utilizing multimodal information from human speech. Previous studies have focused on incorporating additional modalities to enhance the quality of generated gestures. However, these methods perform poorly when certain modalities are missing during inference. To address this problem, we suggest using speech-derived multimodal priors to improve gesture generation. We introduce a novel method that separates priors from speech and employs multimodal priors as constraints for generating gestures. Our approach utilizes a chain-like modeling method to generate facial blendshapes, body movements, and hand gestures sequentially. Specifically, we incorporate rhythm cues derived from facial deformation and stylization prior based on speech emotions, into the process of generating gestures. By incorporating multimodal priors, our method improves the quality of generated gestures and eliminate the need for expensive setup preparation during inference. Extensive experiments and user studies confirm that our proposed approach achieves state-of-the-art performance.

Speech-Driven Personalized Gesture Synthetics: Harnessing Automatic Fuzzy Feature Inference.

Speech-driven gesture generation is an emerging field within virtual human creation. However, a significant challenge lies in accurately determining and processing the multitude of input features (such as acoustic, semantic, emotional, personality, and even subtle unknown features). Traditional approaches, reliant on various explicit feature inputs and complex multimodal processing, constrain the expressiveness of resulting gestures and limit their applicability. To address these challenges, we present Persona-Gestor, a novel end-to-end generative model designed to generate highly personalized 3D full-body gestures solely relying on raw speech audio. The model combines a fuzzy feature extractor and a non-autoregressive Adaptive Layer Normalization (AdaLN) transformer diffusion architecture (DiTs-based). The fuzzy feature extractor harnesses a fuzzy inference strategy that automatically infers implicit, continuous fuzzy features. These fuzzy features, represented as a unified latent feature, are fed into the AdaLN transformer. The AdaLN transformer introduces a conditional mechanism that applies a uniform function across all tokens, thereby effectively modeling the correlation between the fuzzy features and the gesture sequence. This module ensures a high level of gesture-speech synchronization while preserving naturalness. Finally, we employ the diffusion model to train and infer various gestures. Extensive subjective and objective evaluations on the Trinity, ZEGGS, and BEAT datasets confirm our model's superior performance to the current state-of-the-art approaches. Persona-Gestor improves the system's usability and generalization capabilities, setting a new benchmark in speech-driven gesture synthesis and broadening the horizon for virtual human technology.

Cross-Modal Quantization for Co-Speech Gesture Generation

Cross-Modal Quantization for Co-Speech Gesture Generation

It takes two, not one: context-aware nonverbal behaviour generation in dyadic interactions

Nonverbal behaviours are integral parts of human social interaction. Equipping social robots with human nonverbal communication skills has been an active research area for decades, where data-driven, end-to-end learning approaches have become predominant in recent years, offering scalability and generalisability. However, most of the current works only consider social signals of a single character to model co-speech gestures in non-interactive settings. To address this shortcoming, this paper introduces a context-aware Generative Adversarial Network, intending to produce social cues for robots. The approach captures both intra- and interpersonal social signals of two interlocutors to model body gestures in dyadic interaction. We conducted a series of experiments to validate the proposed solution under different interaction settings. First, the experimental results conducted in the JESTKOD dataset demonstrate the contribution of encoding context, namely the behaviours of the interaction partner, in the prediction of target person's gestures in agreement situations. Second, the experiments conducted in the new LISI-HHI dataset show that combining Discriminator and Context Encoder results in a gesture generation framework that is effective across various social communication contexts.

Verbal and nonverbal fluency in amyotrophic lateral sclerosis.

Amyotrophic lateral sclerosis (ALS) is a multi-system disorder that commonly affects cognition and behaviour. Verbal fluency impairments are consistently reported in ALS patients, and we aimed to investigate whether this deficit extends beyond the verbal domain. We further aimed to determine whether deficits are underpinned by a primary intrinsic response generation impairment (i.e., a global reduction across tasks), potentially related to apathy, or an inability to maintain responding over time (i.e., a 'drop off' pattern). Twenty-two ALS patients and 21 demographically-matched controls completed verbal and nonverbal fluency tasks (phonemic/semantic word fluency, design fluency, gesture fluency and ideational fluency), requiring the generation of responses over a specified time period. Fluency performance was analysed in terms of the overall number of novel items produced, as well as the number of items produced in the first 'initiation' and the remaining 'maintenance' time periods. ALS patients' overall performance was not globally reduced across tasks. Patients were impaired only on meaningful gesture fluency, which requires the generation of gestures that communicate meaning (e.g., waving). On phonemic fluency, ALS patients showed a 'drop off' pattern of performance, where they had difficulty maintaining responding over time, but this pattern was not evident on the other fluency tasks. Apathy did not appear to be related to fluency performance. The selective meaningful gesture fluency deficit, in the context of preserved meaningless gesture fluency, highlights that the retrieval of action knowledge may be weakened in early ALS.

Extrovert or Introvert? GAN-Based Humanoid Upper-Body Gesture Generation for Different Impressions

AbstractGestures, a form of body language, significantly influence how users perceive humanoid robots. Recent data-driven methods for co-speech gestures have successfully enhanced the naturalness of the generated gestures. Moreover, compared to rule-based systems, these methods are more generalizable for unseen speech input. However, many of these methods cannot directly influence people’s perceptions of robots. The primary challenge lies in the intricacy of constructing a dataset with varied impression labels to develop a conditional generation model. In our prior work ([22]) Controlling the impression of robots via gan-based gesture generation. In:Proceedings of the international conference on intelligent robots and systems. IEEE, pp 9288-9295), we introduced a heuristic approach for automatic labeling, training a deep learning model to control robot impressions. We demonstrated the model’s effectiveness on both a virtual agent and a humanoid robot. In this study, we refined the motion retargeting algorithm for the humanoid robot and conducted a user study using four questions representing different aspects of extroversion. Our results show an improved capability in controlling the perceived degree of extroversion in the humanoid robot compared to previous methods. Furthermore, we discovered that different aspects of extroversion interact uniquely with motion statistics

ACT2G

Recent increase of remote-work, online meeting and tele-operation task makes people find that gesture for avatars and communication robots is more important than we have thought. It is one of the key factors to achieve smooth and natural communication between humans and AI systems and has been intensively researched. Current gesture generation methods are mostly based on deep neural network using text, audio and other information as the input, however, they generate gestures mainly based on audio, which is called a beat gesture. Although the ratio of the beat gesture is more than 70% of actual human gestures, content based gestures sometimes play an important role to make avatars more realistic and human-like. In this paper, we propose a attention-based contrastive learning for text-to-gesture (ACT2G), where generated gestures represent content of the text by estimating attention weight for each word from the input text. In the method, since text and gesture features calculated by the attention weight are mapped to the same latent space by contrastive learning, once text is given as input, the network outputs a feature vector which can be used to generate gestures related to the content. User study confirmed that the gestures generated by ACT2G were better than existing methods. In addition, it was demonstrated that wide variation of gestures were generated from the same text by changing attention weights by creators.

GestureDiffuCLIP: Gesture Diffusion Model with CLIP Latents

The automatic generation of stylized co-speech gestures has recently received increasing attention. Previous systems typically allow style control via predefined text labels or example motion clips, which are often not flexible enough to convey user intent accurately. In this work, we present GestureDiffuCLIP, a neural network framework for synthesizing realistic, stylized co-speech gestures with flexible style control. We leverage the power of the large-scale Contrastive-Language-Image-Pre-training (CLIP) model and present a novel CLIP-guided mechanism that extracts efficient style representations from multiple input modalities, such as a piece of text, an example motion clip, or a video. Our system learns a latent diffusion model to generate high-quality gestures and infuses the CLIP representations of style into the generator via an adaptive instance normalization (AdaIN) layer. We further devise a gesture-transcript alignment mechanism that ensures a semantically correct gesture generation based on contrastive learning. Our system can also be extended to allow fine-grained style control of individual body parts. We demonstrate an extensive set of examples showing the flexibility and generalizability of our model to a variety of style descriptions. In a user study, we show that our system outperforms the state-of-the-art approaches regarding human likeness, appropriateness, and style correctness.

Zero-shot style transfer for gesture animation driven by text and speech using adversarial disentanglement of multimodal style encoding.

Modeling virtual agents with behavior style is one factor for personalizing human-agent interaction. We propose an efficient yet effective machine learning approach to synthesize gestures driven by prosodic features and text in the style of different speakers including those unseen during training. Our model performs zero-shot multimodal style transfer driven by multimodal data from the PATS database containing videos of various speakers. We view style as being pervasive; while speaking, it colors the communicative behaviors expressivity while speech content is carried by multimodal signals and text. This disentanglement scheme of content and style allows us to directly infer the style embedding even of a speaker whose data are not part of the training phase, without requiring any further training or fine-tuning. The first goal of our model is to generate the gestures of a source speaker based on the content of two input modalities-Mel spectrogram and text semantics. The second goal is to condition the source speaker's predicted gestures on the multimodal behavior style embedding of a target speaker. The third goal is to allow zero-shot style transfer of speakers unseen during training without re-training the model. Our system consists of two main components: (1) a speaker style encoder network that learns to generate a fixed-dimensional speaker embedding style from a target speaker multimodal data (mel-spectrogram, pose, and text) and (2) a sequence-to-sequence synthesis network that synthesizes gestures based on the content of the input modalities-text and mel-spectrogram-of a source speaker and conditioned on the speaker style embedding. We evaluate that our model is able to synthesize gestures of a source speaker given the two input modalities and transfer the knowledge of target speaker style variability learned by the speaker style encoder to the gesture generation task in a zero-shot setup, indicating that the model has learned a high-quality speaker representation. We conduct objective and subjective evaluations to validate our approach and compare it with baselines.

A Comprehensive Review of Data‐Driven Co‐Speech Gesture Generation

AbstractGestures that accompany speech are an essential part of natural and efficient embodied human communication. The automatic generation of such co‐speech gestures is a long‐standing problem in computer animation and is considered an enabling technology for creating believable characters in film, games, and virtual social spaces, as well as for interaction with social robots. The problem is made challenging by the idiosyncratic and non‐periodic nature of human co‐speech gesture motion, and by the great diversity of communicative functions that gestures encompass. The field of gesture generation has seen surging interest in the last few years, owing to the emergence of more and larger datasets of human gesture motion, combined with strides in deep‐learning‐based generative models that benefit from the growing availability of data. This review article summarizes co‐speech gesture generation research, with a particular focus on deep generative models. First, we articulate the theory describing human gesticulation and how it complements speech. Next, we briefly discuss rule‐based and classical statistical gesture synthesis, before delving into deep learning approaches. We employ the choice of input modalities as an organizing principle, examining systems that generate gestures from audio, text and non‐linguistic input. Concurrent with the exposition of deep learning approaches, we chronicle the evolution of the related training data sets in terms of size, diversity, motion quality, and collection method (e.g., optical motion capture or pose estimation from video). Finally, we identify key research challenges in gesture generation, including data availability and quality; producing human‐like motion; grounding the gesture in the co‐occurring speech in interaction with other speakers, and in the environment; performing gesture evaluation; and integration of gesture synthesis into applications. We highlight recent approaches to tackling the various key challenges, as well as the limitations of these approaches, and point toward areas of future development.