Minhao Liu

Yue Zhang, Zhizhou Zhong, Minhao Liu et al.

Real-time video dubbing that preserves identity consistency while achieving accurate lip synchronization remains a critical challenge. Existing approaches face a trilemma: diffusion-based methods achieve high visual fidelity but suffer from prohibitive computational costs, while GAN-based solutions sacrifice lip-sync accuracy or dental details for real-time performance. We present MuseTalk, a novel two-stage training framework that resolves this trade-off through latent space optimization and spatio-temporal data sampling strategy. Our key innovations include: (1) During the Facial Abstract Pretraining stage, we propose Informative Frame Sampling to temporally align reference-source pose pairs, eliminating redundant feature interference while preserving identity cues. (2) In the Lip-Sync Adversarial Finetuning stage, we employ Dynamic Margin Sampling to spatially select the most suitable lip-movement-promoting regions, balancing audio-visual synchronization and dental clarity. (3) MuseTalk establishes an effective audio-visual feature fusion framework in the latent space, delivering 30 FPS output at 256*256 resolution on an NVIDIA V100 GPU. Extensive experiments demonstrate that MuseTalk outperforms state-of-the-art methods in visual fidelity while achieving comparable lip-sync accuracy. %The codes and models will be made publicly available upon acceptance. The code is made available at \href{https://github.com/TMElyralab/MuseTalk}{https://github.com/TMElyralab/MuseTalk}

Ailing Zeng, Xiao Sun, Lei Yang et al.

Various deep learning techniques have been proposed to solve the single-view 2D-to-3D pose estimation problem. While the average prediction accuracy has been improved significantly over the years, the performance on hard poses with depth ambiguity, self-occlusion, and complex or rare poses is still far from satisfactory. In this work, we target these hard poses and present a novel skeletal GNN learning solution. To be specific, we propose a hop-aware hierarchical channel-squeezing fusion layer to effectively extract relevant information from neighboring nodes while suppressing undesired noises in GNN learning. In addition, we propose a temporal-aware dynamic graph construction procedure that is robust and effective for 3D pose estimation. Experimental results on the Human3.6M dataset show that our solution achieves 10.3\% average prediction accuracy improvement and greatly improves on hard poses over state-of-the-art techniques. We further apply the proposed technique on the skeleton-based action recognition task and also achieve state-of-the-art performance. Our code is available at https://github.com/ailingzengzzz/Skeletal-GNN.

Qiuxia Lai, Yu Li, Ailing Zeng et al.

The selective visual attention mechanism in the human visual system (HVS) restricts the amount of information to reach visual awareness for perceiving natural scenes, allowing near real-time information processing with limited computational capacity [Koch and Ullman, 1987]. This kind of selectivity acts as an 'Information Bottleneck (IB)', which seeks a trade-off between information compression and predictive accuracy. However, such information constraints are rarely explored in the attention mechanism for deep neural networks (DNNs). In this paper, we propose an IB-inspired spatial attention module for DNN structures built for visual recognition. The module takes as input an intermediate representation of the input image, and outputs a variational 2D attention map that minimizes the mutual information (MI) between the attention-modulated representation and the input, while maximizing the MI between the attention-modulated representation and the task label. To further restrict the information bypassed by the attention map, we quantize the continuous attention scores to a set of learnable anchor values during training. Extensive experiments show that the proposed IB-inspired spatial attention mechanism can yield attention maps that neatly highlight the regions of interest while suppressing backgrounds, and bootstrap standard DNN structures for visual recognition tasks (e.g., image classification, fine-grained recognition, cross-domain classification). The attention maps are interpretable for the decision making of the DNNs as verified in the experiments. Our code is available at https://github.com/ashleylqx/AIB.git.

Ailing Zeng, Minhao Liu, Zhiwei Liu et al.

We propose a novel solution to addressing a long-standing dilemma in the representation learning of graph neural networks (GNNs): how to effectively capture and represent useful information embedded in long-distance nodes to improve the performance of nodes with low homophily without leading to performance degradation in nodes with high homophily. This dilemma limits the generalization capability of existing GNNs. Intuitively, interactions with distant nodes introduce more noise for a node than those with close neighbors. However, in most existing works, messages being passed among nodes are mingled together, which is inefficient from a communication perspective. Our solution is based on a novel, simple, yet effective aggregation scheme, resulting in a ladder-style GNN architecture, namely LADDER-GNN. Specifically, we separate messages from different hops, assign different dimensions for them, and then concatenate them to obtain node representations. Such disentangled representations facilitate improving the information-to-noise ratio of messages passed from different hops. To explore an effective hop-dimension relationship, we develop a conditionally progressive neural architecture search strategy. Based on the searching results, we further propose an efficient approximate hop-dimension relation function to facilitate the rapid configuration of the proposed LADDER-GNN. We verify the proposed LADDER-GNN on seven diverse semi-supervised node classification datasets. Experimental results show that our solution achieves better performance than most existing GNNs. We further analyze our aggregation scheme with two commonly used GNN architectures, and the results corroborate that our scheme outperforms existing schemes in classifying low homophily nodes by a large margin.

Ailing Zeng, Xiao Sun, Fuyang Huang et al.

Human poses that are rare or unseen in a training set are challenging for a network to predict. Similar to the long-tailed distribution problem in visual recognition, the small number of examples for such poses limits the ability of networks to model them. Interestingly, local pose distributions suffer less from the long-tail problem, i.e., local joint configurations within a rare pose may appear within other poses in the training set, making them less rare. We propose to take advantage of this fact for better generalization to rare and unseen poses. To be specific, our method splits the body into local regions and processes them in separate network branches, utilizing the property that a joint position depends mainly on the joints within its local body region. Global coherence is maintained by recombining the global context from the rest of the body into each branch as a low-dimensional vector. With the reduced dimensionality of less relevant body areas, the training set distribution within network branches more closely reflects the statistics of local poses instead of global body poses, without sacrificing information important for joint inference. The proposed split-and-recombine approach, called SRNet, can be easily adapted to both single-image and temporal models, and it leads to appreciable improvements in the prediction of rare and unseen poses.

Fuyang Huang, Ailing Zeng, Minhao Liu et al.

In this paper, we propose a two-stage fully 3D network, namely \textbf{DeepFuse}, to estimate human pose in 3D space by fusing body-worn Inertial Measurement Unit (IMU) data and multi-view images deeply. The first stage is designed for pure vision estimation. To preserve data primitiveness of multi-view inputs, the vision stage uses multi-channel volume as data representation and 3D soft-argmax as activation layer. The second one is the IMU refinement stage which introduces an IMU-bone layer to fuse the IMU and vision data earlier at data level. without requiring a given skeleton model a priori, we can achieve a mean joint error of $28.9$mm on TotalCapture dataset and $13.4$mm on Human3.6M dataset under protocol 1, improving the SOTA result by a large margin. Finally, we discuss the effectiveness of a fully 3D network for 3D pose estimation experimentally which may benefit future research.

Fuyang Huang, Ailing Zeng, Minhao Liu et al.

In this paper, we propose a novel structure-aware 3D hourglass network for hand pose estimation from a single depth image, which achieves state-of-the-art results on MSRA and NYU datasets. Compared to existing works that perform image-to-coordination regression, our network takes 3D voxel as input and directly regresses 3D heatmap for each joint. To be specific, we use hourglass network as our backbone network and modify it into 3D form. We explicitly model tree-like finger bone into the network as well as in the loss function in an end-to-end manner, in order to take the skeleton constraints into consideration. Final estimation can then be easily obtained from voxel density map with simple post-processing. Experimental results show that the proposed structure-aware 3D hourglass network is able to achieve a mean joint error of 7.4 mm in MSRA and 8.9 mm in NYU datasets, respectively.