Although audio-driven talking face generation has witnessed significant advances in recent years, two problems remain to be solved. First, the existing models cannot control the long-term head actions as humans expect, because the audio can only provide short-term cues, such as the rhythms and sentiments, for head movements. Second, generating long-term head poses and ensuring accurate lip motions remain challenging due to the difficulty in harnessing the optimization process for large-scale head movements and small-scale mouth motions. In this study, we propose a novel method to address these issues. First, to alleviate the limitations of audio conditions, we propose a Pose Latent Diffusion (PLD) model to generate head motions from two kinds of input modalities: the input audio and user-controlled text prompts. The audio provides short-term rhythm correspondence with the head movements, while the text prompts describe the long-term semantics of head motions. Second, we propose a refinement-based learning strategy to synthesize head movements and accurate lip motions using two cascaded networks, namely CoarseNet and RefineNet. The CoarseNet estimates coarse global motions to produce animated images with changed poses, and the RefineNet progressively estimates finer lip motions from low to high resolutions, yielding improved lip-synchronization performance. Experiments demonstrate that our method can achieve better pose diversity and realness compared to audio-based pose generation baselines, and our video generator model outperforms state-of-the-art methods in synthesizing natural head motions. Projects and demos are available at https://junleen.github.io/projects/posetalk .

With the rapid development of social media, the amount of face video data has grown rapidly, making face video compression a hot research topic. Traditional video coding techniques do not discriminate video content and compress all videos in the same way, while talking head video compression should have more potential. Existing generative compression methods mostly adopt static reference frames, resulting in a decrease in fidelity caused by dynamic background or large pose change. In this article, we propose a hybrid compression scheme for face videos which combines traditional coding with generative compression. On the one hand, we sample and encode key frames with traditional codecs to provide dynamic reference frames which contain real-time background and motion information. On the other hand, we devise a deep video generation model to synthesize smooth video frames according to the extracted sparse keypoints. Combining the pixel-level recovery capability of traditional coding with the detail generation capability of deep generative models, our proposed hybrid scheme is able to implement high-fidelity face video compression at low bitrate in real time. Additionally, we also devise a Portrait Recovery module to recover the low-quality key frames, improving the reconstruction quality in low-bitrate scenarios. Extensive experiments show that our method has advantages over traditional codecs and existing generative compression methods in terms of both rate-distortion performance and coding complexity.

Despite the progress of 3D Gaussian Splatting (3DGS), reconstructing one-shot animatable 3D human avatars from a single image remains a challenging task. Existing 3DGS-based methods primarily rely on appearance observation and motion cues from monocular videos to reconstruct animatable 3D avatars. However, when applied to single-image setups, these methods struggle to extract accurate 3D features from a 2D image, limiting their ability to capture fine-grained appearance details and dynamic deformation, especially from challenging perspectives. In this work, we propose SIAvatar, a novel single-image human reconstruction method that integrates diffusion-based appearance prior and parametric human model geometry prior within a 3DGS framework. Secondly, a vertex-based adaptive Gaussian densification scheme is introduced to effectively represent human geometry while mitigating artifacts. Extensive experiments demonstrate that SIAvatar generates realistic 3D avatars with plausible appearance details and novel pose animation from a single input image. Project page: https://siavatar.github.io/.

3D digital garment generation and editing play a pivotal role in fashion design, virtual try-on, and gaming. Traditional methods struggle to meet the growing demand due to technical complexity and high resource costs. Learning-based approaches offer faster, more diverse garment synthesis based on specific requirements and reduce human efforts and time costs. However, they still face challenges such as inconsistent multi-view geometry or textures and heavy reliance on detailed garment topology and manual rigging. We propose SemanticGarment, a 3D Gaussian-based method that realizes high-fidelity 3D garment generation from text or image prompts and supports semantic-based interactive editing for flexible user customization. To ensure multi-view consistency and garment fitting, we propose to leverage structural human priors for the generative model by introducing a 3D semantic clothing model, which initializes the geometry structure and lays the groundwork for view-consistent garment generation and editing. Without the need to regenerate or rely on existing mesh templates, our approach allows for rapid and diverse modifications to existing Gaussians, either globally or within a local region. To address the artifacts caused by self-occlusion for garment reconstruction based on single image, we develop a self-occlusion optimization strategy to mitigate holes and artifacts that arise when directly animating self-occluded garments. Extensive experiments are conducted to demonstrate our superior performance in 3D garment generation and editing.

In this work, we address the task of table image to LaTeX code generation, with the goal of automating the reconstruction of high-quality, publication-ready tables from visual inputs. A central challenge of this task lies in accurately handling complex tables—those with large sizes, deeply nested structures, and semantically rich or irregular cell content—where existing methods often fail. We begin with a comprehensive analysis, identifying key challenges and highlighting the limitations of current evaluation protocols. To overcome these issues, we propose a reinforced multimodal large language model (MLLM) framework, where a pre-trained MLLM is fine-tuned on a large-scale table-to-LaTeX dataset. To further improve generation quality, we introduce a dual-reward reinforcement learning strategy based on Group Relative Policy Optimization (GRPO). Unlike standard approaches that optimize purely over text outputs, our method incorporates both a structure-level reward on LaTeX code and a visual fidelity reward computed from rendered outputs, enabling direct optimization of the visual output quality. We adopt a hybrid evaluation protocol combining TEDS-Structure and CW-SSIM, and show that our method achieves state-of-the-art performance, particularly on structurally complex tables, demonstrating the effectiveness and robustness of our approach.

ABSTRACT Synthesis of 3D‐aware facial images from latent spaces has garnered significant attention in multimedia content generation due to its ability to model images with rich semantics and diverse appearances. However, existing methods often rely on labeled data or suffer from incomplete attribute control and ambiguous latent space semantics. This paper proposes an efficient semantic distillation method that learns attribute directions of pre‐trained 3D GAN models, without the supervised semantic labels. We consider the latent space of GAN models as the mixture of two featured subspaces, namely the geometry‐aware space and appearance‐aware space. Following this hypothesis, we define two sets of learnable latent bases and use linear composition to represent controllable geometry and appearance feature space, respectively. To learn semantic‐wise latent bases for attribute‐controllable image generation, we design a framework and propose a three‐staged training strategy, which optimizes the appearance‐aware and the geometry‐aware latent bases. With the two sets of latent bases, we obtain the combined latent vectors using different weights for those bases and synthesize images with specified attributes. Compared to existing methods, our approach eliminates the need for labeled data and enables more controllable attribute disentanglement while ensuring identity consistency, which can be directly applied to real‐world scenarios such as virtual avatars and augmented reality applications. Experiments demonstrate the effectiveness and insight of our approach in aiding a better understanding of the latent space of 3D GANs.

With the rapid expansion of smart cities, Intelligent Transportation Systems (ITS) are assuming an increasingly pivotal role. Among the multitude of tasks within ITS, traffic state forecasting stands out. As city boundaries grow, traffic forecasting encounters scalability and network transmission challenges. This research contributes to traffic state forecasting within large-scale, massive Internet of Things (IoT) scenarios. By investigating an urban expressway managing architecture that employs LoRaWAN communication, a novel deep learning-based model named Time Alignment based Temporal-Graph Attention Network (TATGaN) is proposed. Using the temporal-graph attention mechanism, TATGaN is able to extract temporal-spatial information and predict traffic state accurately. Moreover, the time alignment block makes TATGaN capable of handling irregular sequences given by the asynchronous arrival of packets. Simulation results based on OSM data of a specific region within Abu Dhabi show that TATGaN outperforms existing baseline methods in prediction performance with lower error and the higher reliability. Furthermore, the performance evaluation demonstrates the suitability of TATGaN for large-scale traffic network scenarios, attributing its efficiency to the transmission schedule and parameter mechanisms in LoRaWAN networks.

Generating photo-realistic avatars from audio plays an important role in extended reality (XR) and metaverse. In this paper, we lift the input audio from speech to singing, which has been rarely studied. The significant distinction between singing and talking poses great challenges for adapting talking face generation methods to the singing regime. To address this, we propose a high-fidelity singing avatar synthesis method called SingAvatar. Besides the audio, we incorporate vocal conditions involving phonemes and variance to alleviate the ambiguity of learning the singing-to-face mapping. Concretely, we tailor a two-stage pipeline: singing voice synthesis and portrait generation from the synthesized audio and auxiliary vocal conditions. Further, we curate a fine-grained singing head dataset containing singing videos with synchronized audio and accurate vocal conditions. In experiments, SingAvatar outperforms competing methods regarding audio-mouth synchronization, the naturalness of head movements, and controllability over the results. The code and dataset will be made publicly available.

Talking face generation aims at generating photo-realistic video portraits of a target person driven by input audio. According to the nature of audio to lip motions mapping, the same speech content may have different appearances even for the same person at different occasions. Such one-to-many mapping problem brings ambiguity during training and thus causes inferior visual results. Although this one-to-many mapping could be alleviated in part by a two-stage framework (i.e., an audio-to-expression model followed by a neural-rendering model), it is still insufficient since the prediction is produced without enough information (e.g., emotions, wrinkles, etc.). In this paper, we propose MemFace to complement the missing information with an implicit memory and an explicit memory that follow the sense of the two stages respectively. More specifically, the implicit memory is employed in the audio-to-expression model to capture high-level semantics in the audio-expression shared space, while the explicit memory is employed in the neural-rendering model to help synthesize pixel-level details. Our experimental results show that our proposed MemFace surpasses all the state-of-the-art results across multiple scenarios consistently and significantly.

Unsupervised face reenactment aims to animate a source image to imitate the motions of a target image while retaining the source portrait’s attributes like facial geometry, identity, hair texture, and background. While prior methods can extract the motion from the target image via compact representations (e.g., keypoints or latent motion bases [ 50 ]), they are not robust in predicting motions that are disentangled with portrait attributes, thus failing to preserve portrait attributes in the cross-subject reenactment. In this work, we propose an effective and cost-efficient face reenactment approach to address this issue. Our approach is highlighted by two major strengths. First, based on the theory of latent motion bases, we disentangle the full-head motion into two parts: the transferable motion and preservable motion and then compose the full motion representation using latent motions from the source image and the target image. Second, to optimize and learn disentangled motions, we introduce an efficient training framework, which features two training strategies: (1) a mixture training strategy that encompasses self-reenactment training and cross-subject training for better motion disentanglement and (2) a multi-path training strategy that improves the visual consistency of portrait attributes. Extensive experiments on widely used benchmarks demonstrate that our method exhibits a remarkable generalization ability compared to state-of-the-art baselines. Project and demos are available at https://junleen.github.io/projects/vicoface .

360° panoramas are extensively utilized as environmental light sources in computer graphics. However, capturing a 360° × 180° panorama poses challenges due to the necessity of specialized and costly equipment, and additional human resources. Prior studies develop various learning-based generative methods to synthesize panoramas from a single Narrow Field-of-View (NFoV) image, but they are limited in alterable input patterns, generation quality, and controllability. To address these issues, we propose a novel pipeline called PanoDiff, which efficiently generates complete 360° panoramas using one or more unregistered NFoV images captured from arbitrary angles. Our approach has two primary components to overcome the limitations. Firstly, a two-stage angle prediction module to handle various numbers of NFoV inputs. Secondly, a novel latent diffusion-based panorama generation model uses incomplete panorama and text prompts as control signals and utilizes several geometric augmentation schemes to ensure geometric properties in generated panoramas. Experiments show that PanoDiff achieves state-of-the-art panoramic generation quality and high controllability, making it suitable for applications such as content editing.

Talking-head video editing aims to efficiently insert, delete, and substitute the word of a pre-recorded video through a text transcript editor. The key challenge for this task is obtaining an editing model that generates new talking-head video clips which simultaneously have accurate lip synchronization and motion smoothness. Previous approaches, including 3DMM-based (3D Morphable Model) methods and NeRF-based (Neural Radiance Field) methods, are sub-optimal in that they either require minutes of source videos and days of training time or lack the disentangled control of verbal (e.g., lip motion) and non-verbal (e.g., head pose and expression) representations for video clip insertion. In this work, we fully utilize the video context to design a novel framework for talking-head video editing, which achieves efficiency, disentangled motion control, and sequential smoothness. Specifically, we decompose this framework to motion prediction and motion-conditioned rendering: (1) We first design an animation prediction module that efficiently obtains smooth and lip-sync motion sequences conditioned on the driven speech. This module adopts a non-autoregressive network to obtain context prior and improve the prediction efficiency, and it learns a speech-animation mapping prior with better generalization to novel speech from a multi-identity video dataset. (2) We then introduce a neural rendering module to synthesize the photo-realistic and full-head video frames given the predicted motion sequence. This module adopts a pre-trained head topology and uses only few frames for efficient fine-tuning to obtain a person-specific rendering model. Extensive experiments demonstrate that our method efficiently achieves smoother editing results with higher image quality and lip accuracy using less data than previous methods.

Face reenactment aims to generate an animation of a source face using the poses and expressions from a target face. Although recent methods have made remarkable progress by exploiting generative adversarial networks, they are limited in generating high-fidelity and identity-preserving results due to the inappropriate driving information and insufficiently effective animating strategies. In this work, we propose a novel face reenactment framework that achieves both high-fidelity generation and identity preservation. Instead of sparse face representations (e.g., facial landmarks and keypoints), we utilize the Projected Normalized Coordinate Code (PNCC) to better preserve facial details. We propose to reconstruct the PNCC with the source identity parameters and the target pose and expression parameters estimated by 3D face reconstruction to factor out the target identity. By adopting the reconstructed representation as the driving information, we address the problem of identity mismatch. To effectively utilize the driving information, we establish the correspondence between the reconstructed representation and the source representation based on the features extracted by an encoder network. This identity-matched correspondence is then utilized to animate the source face using a novel feature transformation strategy. The generator network is further enhanced by the proposed geometry-aware skip connection. Once trained, our model can be applied to previously unseen faces without further training or fine-tuning. Through extensive experiments, we demonstrate the effectiveness of our method in face reenactment and show that our model outperforms state-of-the-art approaches both qualitatively and quantitatively. Additionally, the proposed PNCC reconstruction module can be easily inserted into other methods and improve their performance in cross-identity face reenactment.

We propose a novel robust and efficient Speech-to-Animation (S2A) approach for synchronized facial animation generation in human-computer interaction. Compared with conventional approaches, the proposed approach utilizes phonetic posteriorgrams (PPGs) of spoken phonemes as input to ensure the cross-language and cross-speaker ability, and introduces corresponding prosody features (i.e. pitch and energy) to further enhance the expression of generated animation. Mixture-of-experts (MOE)-based Transformer is employed to better model contextual information while provide significant optimization on computation efficiency. Experiments demonstrate the effectiveness of the proposed approach on both objective and subjective evaluation with 17× inference speedup compared with the state-of-the-art approach.

As the latest video coding standard, versatile video coding (VVC) has shown its ability in retaining pixel quality. To excavate more compression potential for video conference scenarios under ultra-low bitrate, this paper proposes a bitrate-adjustable hybrid compression scheme for face video. This hybrid scheme combines the pixel-level precise recovery capability of traditional coding with the generation capability of deep learning based on abridged information, where Pixel-wise Bi-Prediction, Low-Bitrate-FOM and Lossless Keypoint Encoder collaborate to achieve PSNR up to 36.23 dB at a low bitrate of 1.47 KB/s. Without introducing any additional bi-trate, our method has a clear advantage over VVC under a completely fair comparative experiment, which proves the effectiveness of our proposed scheme. Moreover, our scheme can adapt to any existing encoder/configuration to deal with different encoding requirements, and the bitrate can be dynamically adjusted according to the network condition.

Image composition plays a common but important role in photo editing. To acquire photo-realistic composite images, one must adjust the appearance and visual style of the foreground to be compatible with the background. Existing deep learning methods for harmonizing composite images directly learn an image mapping network from the composite to real one, without explicit exploration on visual style consistency between the background and the foreground images. To ensure the visual style consistency between the foreground and the background, in this paper, we treat image harmonization as a style transfer problem. In particular, we propose a simple yet effective Region-aware Adaptive Instance Normalization (RAIN) module, which explicitly formulates the visual style from the background and adaptively applies them to the foreground. With our settings, our RAIN module can be used as a drop-in module for existing image harmonization networks and is able to bring significant improvements. Extensive experiments on the existing image harmonization benchmark datasets shows the superior capability of the proposed method. Code is available at https://github.com/junleen/RainNet.

In face synthesis tasks, commonly used 2D face representations (e.g. 2D landmarks, segmentation maps, etc.) are usually sparse and discontinuous. To combat these shortcomings, we utilize a dense and continuous representation, named Projected Normalized Coordinate Code (PNCC), as the guidance and develop a PNCC-Spatio-Normalization (PSN) method to achieve face synthesis regarding arbitrary head poses and expressions. Based on PSN, we provide an effective framework for face reenactment and face swapping task. To ensure a harmonious and seamless face swapping, a simple yet effective Appearance-Blending Module (ABM) is proposed to fit the synthesized face to the target face. Our method is subject-agnostic and can be applied to any pair of faces without extra fine-tuning. Both qualitative and quantitative experiments are conducted to demonstrate the superiority of the proposed method in comparisons to existing state-of-the-art systems.

Recent studies have shown remarkable success in synthesizing realistic talking faces by exploiting generative adversarial networks. However, existing methods are mostly target specific that cannot generate images of previously unseen people, and they suffer from artifacts such as blurriness and mismatching of facial details. In this paper, we tackle these problems by proposing a target-agnostic framework. We introduce a geometry-aware feature transformation module to achieve shape transfer while preserving the appearance of the source face. To further improve image quality of synthesized results, we present a multi-scale spatially-consistent transfer unit to maintain spatial consistency between the encoder and decoder features. Experimental results show that our model is able to synthesize photo-realistic talking faces which are previously unseen, outperforming state-of-the-art methods both qualitatively and quantitatively.

We introduce a missing data recovery methodology based on a weighted least squares iterative adaptive approach (IAA). The proposed method is referred to as the missing-data IAA (MIAA) and it can be used for uniform or non-uniform sampling as well as for arbitrary data missing patterns. MIAA uses the IAA spectrum estimates to retrieve the missing data, based on a spectral least squares criterion similar to that used by IAA. Numerical examples are presented to show the effectiveness of MIAA for missing data recovery. We also show that MIAA can outperform an existing competitive approach, and this at a much lower computational cost.

We introduce a missing data recovery methodology based on a weighted least squares iterative adaptive approach (IAA). The proposed method is referred to as the missing-data IAA (MIAA) and it can be used for uniform or nonuniform sampling as well as for arbitrary data missing patterns. MIAA uses the IAA spectrum estimates to retrieve the missing data, by means of either a frequency domain or a time domain approach. Numerical examples are presented to show the effectiveness of MIAA for missing data reconstruction. In particular, we show that MIAA can outperform an existing competitive approach, and this at a much lower computational cost.