In this paper, we propose a novel approach to convert given speech audio to a photo-realistic speaking video of a specific person, where the output video has synchronized, realistic, and expressive rich body dynamics. We achieve this by first generating 3D skeleton movements from the audio sequence using a recurrent neural network (RNN), and then synthesizing the output video via a conditional generative adversarial network (GAN). To make the skeleton movement realistic and expressive, we embed the knowledge of an articulated 3D human skeleton and a learned dictionary of personal speech iconic gestures into the generation process in both learning and testing pipelines. The former prevents the generation of unreasonable body distortion, while the later helps our model quickly learn meaningful body movement through a few recorded videos. To produce photo-realistic and high-resolution video with motion details, we propose to insert part attention mechanisms in the conditional GAN, where each detailed part, e.g. head and hand, is automatically zoomed in to have their own discriminators. To validate our approach, we collect a dataset with 20 high-quality videos from 1 male and 1 female model reading various documents under different topics. Compared with previous SoTA pipelines handling similar tasks, our approach achieves better results by a user study.
DreamWaltz-G: Expressive 3D Gaussian Avatars from Skeleton-Guided 2D Diffusion
The University of Hong Kong (HKU)（香港大学）；Astribot（星动科技）；Tencent（腾讯）；University of Science and Technology of China (USTC)（中国科学技术大学）；International Digital Economy Academy (IDEA)（国际数字经济学院）
Spatial-Temporal Graph Mamba for Music-Guided Dance Video Synthesis
State Key Laboratory of Multimedia Information Processing, School of Computer Science, Peking University（多媒体信息处理国家重点实验室，计算机科学学院，北京大学）；UCAS-Terminus AI Lab, University of Chinese Academy of Sciences（中国科学院大学Terminus AI实验室，中国科学院大学）；School of Intelligent Science and Technology, Nanjing University（智能科学与技术学院，南京大学）；Department of Information Technology and Electrical Engineering, ETH Zurich（信息科技与电气工程系，苏黎世联邦理工学院）；Department of Electrical Engineering, KU Leuven（电气工程系，鲁汶大学）；INSAIT, Sofia Un.（INSAIT，索菲亚大学）；Department of Information Engineering and Computer Science (DISI), University of Trento（信息工程与计算机科学系（DISI），特伦托大学）
Variational Autoencoder (VAE), a simple and effective deep generative model, has led to a number of impressive empirical successes and spawned many advanced variants and theoretical investigations. However, recent studies demonstrate that, when equipped with expressive generative distributions (aka. decoders), VAE suffers from learning uninformative latent representations with the observation called KL Varnishing, in which case VAE collapses into an unconditional generative model. In this work, we introduce mutual posterior-divergence regularization, a novel regularization that is able to cont…
There exist many forms of deep latent variable models, such as the variational autoencoder and adversarial autoencoder. Regardless of the specific class of model, there exists an implicit consensus that the latent distribution should be regularized towards the prior, even in the case where the prior distribution is learned. Upon investigating the effect of latent regularization on image generation our results indicate that in the case where a sufficiently expressive prior is learned, latent regularization is not necessary and may in fact be harmful insofar as image quality is concerned. We add…
Dynamics of human body skeletons convey significant information for human action recognition. Conventional approaches for modeling skeletons usually rely on hand-crafted parts or traversal rules, thus resulting in limited expressive power and difficulties of generalization. In this work, we propose a novel model of dynamic skeletons called Spatial-Temporal Graph Convolutional Networks (ST-GCN), which moves beyond the limitations of previous methods by automatically learning both the spatial and temporal patterns from data. This formulation not only leads to greater expressive power but also st…
Vector Institute for Artificial Intelligence（向量人工智能研究院）；University of Toronto（多伦多大学）
In this paper, we describe the "implicit autoencoder" (IAE), a generative autoencoder in which both the generative path and the recognition path are parametrized by implicit distributions. We use two generative adversarial networks to define the reconstruction and the regularization cost functions of the implicit autoencoder, and derive the learning rules based on maximum-likelihood learning. Using implicit distributions allows us to learn more expressive posterior and conditional likelihood distributions for the autoencoder. Learning an expressive conditional likelihood distribution enables t…
Recent advances in generative adversarial networks (GANs) have opened up the possibility of generating high-resolution photo-realistic images that were impossible to produce previously. The ability of GANs to sample from high-dimensional distributions has naturally motivated researchers to leverage their power for modeling the image prior in inverse problems. We extend this line of research by developing a Bayesian image reconstruction framework that utilizes the full potential of a pre-trained StyleGAN2 generator, which is the currently dominant GAN architecture, for constructing the prior di…
Histopathology image synthesis aims to address the data shortage issue in training deep learning approaches for accurate cancer detection. However, existing methods struggle to produce realistic images that have accurate nuclei boundaries and less artifacts, which limits the application in downstream tasks. To address the challenges, we propose a novel approach that enhances the quality of synthetic images by using nuclei topology and contour regularization. The proposed approach uses the skeleton map of nuclei to integrate nuclei topology and separate touching nuclei. In the loss function, we…
Deep learning-based 3D human pose estimation performs best when trained on large amounts of labeled data, making combined learning from many datasets an important research direction. One obstacle to this endeavor are the different skeleton formats provided by different datasets, i.e., they do not label the same set of anatomical landmarks. There is little prior research on how to best supervise one model with such discrepant labels. We show that simply using separate output heads for different skeletons results in inconsistent depth estimates and insufficient information sharing across skeleto…
Peking University（北京大学）；Shanghai AI Laboratory（上海人工智能实验室）；SenseTime Research（商汤科技研究院）；Southeast University（东南大学）；Nanyang Technological University（南洋理工大学）
We present a novel framework that brings the 3D motion retargeting task from controlled environments to in-the-wild scenarios. In particular, our method is capable of retargeting body motion from a character in a 2D monocular video to a 3D character without using any motion capture system or 3D reconstruction procedure. It is designed to leverage massive online videos for unsupervised training, needless of 3D annotations or motion-body pairing information. The proposed method is built upon two novel canonicalization operations, structure canonicalization and view canonicalization. Trained with…
DreaMo: Articulated 3D Reconstruction From A Single Casual Video
National Tsing Hua University（国立清华大学）；Carnegie Mellon University（卡内基梅隆大学）；University of California, Merced（加州大学默塞德分校）；University of Illinois Urbana-Champaign（伊利诺伊大学厄巴纳-香槟分校）；Amazon（亚马逊）
针对单段随手拍摄互联网视频视角不全的问题，本文提出 DreaMo，结合视角条件扩散先验、定制正则化及骨架生成策略，实现高质量铰接式 3D 形状与可解释骨架重建。
