Mengjian Li
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Mengjian Li has a total of 31 co-authors. Here are the top co-authors by number of shared publications:
- Yaxiong Wang
Hefei University of Technology, Hefei, China
4 shared publications - Fan Zhang
Communication University of Zhejiang, Hangzhou, China
3 shared publications - Fuxing Gao
College of Media Engineering Communication University of Zhejiang China
3 shared publications - Hao Wu
College of Media Engineering, Communication University of Zhejiang, Hangzhou, China
3 shared publications - Hui Du
College of Media Engineering, Communication University of Zhejiang, Hangzhou, China
3 shared publications - Naye Ji
Communication University of Zhejiang
3 shared publications
Latest Publications
- Open Access
- 9 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 16316-16324
HyperSketch: Controllable Video Sketching in a Style Hyperspace
- 9 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 16316-16324
Open AccessHyperSketch: Controllable Video Sketching in a Style Hyperspace
- Open Access
- 10 June 2026
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Multimedia Computing, Communications, and Applications
- Vol. 22 (8) , 1-17
POS: A Prompts Optimization Suite for Augmenting Text-to-Video Generation
This paper targets to enhance the diffusion-based text-to-video generation by improving the two input prompts, including the noise and the text. Accommodated with this goal, we propose POS, a P rompt O ptimization S uite to boost text-to-video models. POS is motivated by two observations: (1) Video generation shows instability in terms of noise. Given the same text, different noises lead to videos that differ significantly in terms of both frame quality and temporal consistency. This observation implies that there exists an optimal noise matched to each textual input; To capture the potential noise, we propose an optimal noise approximator to approach the potential optimal noise. Particularly, the optimal noise approximator initially searches a video that closely relates to the text prompt and then inverts it into the noise space to serve as an improved noise prompt for the textual input. (2) Improving the text prompt via LLMs often causes semantic deviation. Many existing text-to-vision works have utilized LLMs to improve the text prompts for generation enhancement.However, existing methods often neglect the semantic alignment between the original text and the rewritten one.In response to this issue, we design a semantic-preserving rewriter to impose constraints in both rewriting and denoising phrases to preserve semantic consistency. Extensive experiments on popular benchmarks show that our POS can improve the text-to-video models with a clear margin.
- 10 June 2026
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Multimedia Computing, Communications, and Applications
- Vol. 22 (8) , 1-17
Open AccessPOS: A Prompts Optimization Suite for Augmenting Text-to-Video Generation
This paper targets to enhance the diffusion-based text-to-video generation by improving the two input prompts, including the noise and the text. Accommodated with this goal, we propose POS, a P rompt O ptimization S uite to boost text-to-video models. POS is motivated by two observations: (1) Video generation shows instability in terms of noise. Given the same text, different noises lead to videos that differ significantly in terms of both frame quality and temporal consistency. This observation implies that there exists an optimal noise matched to each textual input; To capture the potential noise, we propose an optimal noise approximator to approach the potential optimal noise. Particularly, the optimal noise approximator initially searches a video that closely relates to the text prompt and then inverts it into the noise space to serve as an improved noise prompt for the textual input. (2) Improving the text prompt via LLMs often causes semantic deviation. Many existing text-to-vision works have utilized LLMs to improve the text prompts for generation enhancement.However, existing methods often neglect the semantic alignment between the original text and the rewritten one.In response to this issue, we design a semantic-preserving rewriter to impose constraints in both rewriting and denoising phrases to preserve semantic consistency. Extensive experiments on popular benchmarks show that our POS can improve the text-to-video models with a clear margin.
- 21 April 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Circuits and Systems for Video Technology
- Vol. PP (10518215) , 1
3 × 3 Kernel is ALL You Need for Vision
Most modern Convolutional Neural Networks (CNNs) employ a multi-branch structure with various-sized convolutions to capture long- and short-range dependencies. However, these CNNs use large kernel convolutions (e.g., astonishingly 101 kernels) and specialized techniques (e.g., reparameterization and sparsity), increasing complexity in both training and inference stages. This paper focuses on designing an efficient CNN based on pure 3×3 convolutions without introducing complex operations and techniques. Specifically, we propose a Spatial Pyramid (SP) block, which consists of the Multi-branch Residual (MbR) module and the Gated-branch Residual (GbR) module. The MbR introduces multiscale pooling as the key component, thus capturing long-range visual cues through large down-sampling rates and shorter-range dependencies through low down-sampling rates while maintaining low computational complexity. Besides, the GbR uses one 3×3 convolution to refine dependencies along spatial and channel dimensions. Based on the SP block, we construct the Spatial Pyramid CNN (SPCNN), a model composed exclusively of Point-Wise Convolution and 3×3 Depth-Wise Convolution. Under comparable computational complexity, SPCNN significantly outperforms the state-of-the-art CNN PeLK (83.6% vs 82.6%) with only 3 × 3 kernels (compared to 101 × 101 kernels in PeLK). Besides, our SPCNN demonstrates comparability with state-of-the-art backbones in lightweight models, object detection, instance segmentation, and semantic segmentation. Moreover, evaluations of four image retrieval benchmarks also demonstrate the effectiveness. All codes are released at https://github.com/xiaolai-sqlai/SPCNN.
- 21 April 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Circuits and Systems for Video Technology
- Vol. PP (10518215) , 1
3 × 3 Kernel is ALL You Need for Vision
Most modern Convolutional Neural Networks (CNNs) employ a multi-branch structure with various-sized convolutions to capture long- and short-range dependencies. However, these CNNs use large kernel convolutions (e.g., astonishingly 101 kernels) and specialized techniques (e.g., reparameterization and sparsity), increasing complexity in both training and inference stages. This paper focuses on designing an efficient CNN based on pure 3×3 convolutions without introducing complex operations and techniques. Specifically, we propose a Spatial Pyramid (SP) block, which consists of the Multi-branch Residual (MbR) module and the Gated-branch Residual (GbR) module. The MbR introduces multiscale pooling as the key component, thus capturing long-range visual cues through large down-sampling rates and shorter-range dependencies through low down-sampling rates while maintaining low computational complexity. Besides, the GbR uses one 3×3 convolution to refine dependencies along spatial and channel dimensions. Based on the SP block, we construct the Spatial Pyramid CNN (SPCNN), a model composed exclusively of Point-Wise Convolution and 3×3 Depth-Wise Convolution. Under comparable computational complexity, SPCNN significantly outperforms the state-of-the-art CNN PeLK (83.6% vs 82.6%) with only 3 × 3 kernels (compared to 101 × 101 kernels in PeLK). Besides, our SPCNN demonstrates comparability with state-of-the-art backbones in lightweight models, object detection, instance segmentation, and semantic segmentation. Moreover, evaluations of four image retrieval benchmarks also demonstrate the effectiveness. All codes are released at https://github.com/xiaolai-sqlai/SPCNN.
- Open Access
- 3 August 2024
- Published by Elsevier in SSRN Electronic Journal
Towards Geometric-Photometric Joint Alignment for Facial Mesh Registration
This paper presents a Geometric-Photometric Joint Alignment(GPJA) method, which aligns discrete human expressions at pixel-level by combining geometric and photometric information. Common practices for registering human heads typically involve aligning landmarks with facial template meshes using geometry processing approaches, but often overlook dense pixel-level photometric consistency. This oversight leads to inconsistent texture parameterization across different expressions, hindering the creation of topologically consistent head meshes widely used in movies and games. GPJA overcomes this limitation by leveraging differentiable rendering to align vertices with target expressions, achieving joint alignment in both geometry and photometric appearances automatically, without requiring semantic annotation or pre-aligned meshes for training. It features a holistic rendering alignment strategy and a multiscale regularized optimization for robust and fast convergence on large-step deformation. The method utilizes derivatives at vertex positions for supervision and employs a gradient-based algorithm which guarantees smoothness and avoids topological defects during the geometry evolution. Experimental results demonstrate faithful alignment under various expressions, surpassing the conventional non-rigid ICP-based methods and the state-of-the-art deep learning based method. In practical, our semantic annotation-free method enhances the efficiency of obtaining topology-consistent face models from multi-view stereo facial scanning.
- 3 August 2024
- Published by Elsevier in SSRN Electronic Journal
Open AccessTowards Geometric-Photometric Joint Alignment for Facial Mesh Registration
This paper presents a Geometric-Photometric Joint Alignment(GPJA) method, which aligns discrete human expressions at pixel-level by combining geometric and photometric information. Common practices for registering human heads typically involve aligning landmarks with facial template meshes using geometry processing approaches, but often overlook dense pixel-level photometric consistency. This oversight leads to inconsistent texture parameterization across different expressions, hindering the creation of topologically consistent head meshes widely used in movies and games. GPJA overcomes this limitation by leveraging differentiable rendering to align vertices with target expressions, achieving joint alignment in both geometry and photometric appearances automatically, without requiring semantic annotation or pre-aligned meshes for training. It features a holistic rendering alignment strategy and a multiscale regularized optimization for robust and fast convergence on large-step deformation. The method utilizes derivatives at vertex positions for supervision and employs a gradient-based algorithm which guarantees smoothness and avoids topological defects during the geometry evolution. Experimental results demonstrate faithful alignment under various expressions, surpassing the conventional non-rigid ICP-based methods and the state-of-the-art deep learning based method. In practical, our semantic annotation-free method enhances the efficiency of obtaining topology-consistent face models from multi-view stereo facial scanning.
- Open Access
- 10 July 2024
- Published by Association for Computing Machinery (ACM)
- p. 2177-2187
CaLa: Complementary Association Learning for Augmenting Comoposed Image Retrieval
Composed image retrieval (CIR) is the task of searching target images using an image-text pair as a query. Given the straightforward relation of query pair-target image, the dominant methods follow the learning paradigm of common image-text retrieval and simply model this problem as the query-target matching problem. Particularly, the common practice first encodes the multi-modal query into one feature and then aligns it with the target image. However, such a learning paradigm only explores the naive relation in the triplets. We argue that CIR triplets encompass additional associations besides the primary query-target relation, which is overlooked in existing works. In this paper, we disclose two new relations residing in the triplets by viewing the triplet as a graph node. In analogy with the graph node, we mine two associations of text-bridged image alignment and complementary text reasoning. The text-bridged image alignment considers composed image retrieval as a specialized form of image retrieval, where the query text acts as a bridge between the query image and the target one, and a hinge-based cross attention is proposed to incorporate this relation into the network learning. On the other hand, the association of complementary text reasoning regards composed image retrieval as a specific type of cross-modal retrieval, where the composite two images are used to reason the complementary text. To integrate these views effectively, a twin attention-based compositor is designed. By combining these two types of complementary associations with the explicit query pair-target image relation, we establish a comprehensive set of constraints for composed image retrieval. With the above designs, we finally developed our CaLa, a Complementary Association Learning framework for Augmenting Composed Image Retrieval. Experimental evaluations are conducted on the widely-used CIRR and FashIionIQ benchmarks with multiple backbones to validate the effectiveness of our CaLa. The results demonstrate the superiority of our method in the composed image retrieval task. Our code and models are available at https://github.com/Chiangsonw/CaLa
- 10 July 2024
- Published by Association for Computing Machinery (ACM)
- p. 2177-2187
Open AccessCaLa: Complementary Association Learning for Augmenting Comoposed Image Retrieval
Composed image retrieval (CIR) is the task of searching target images using an image-text pair as a query. Given the straightforward relation of query pair-target image, the dominant methods follow the learning paradigm of common image-text retrieval and simply model this problem as the query-target matching problem. Particularly, the common practice first encodes the multi-modal query into one feature and then aligns it with the target image. However, such a learning paradigm only explores the naive relation in the triplets. We argue that CIR triplets encompass additional associations besides the primary query-target relation, which is overlooked in existing works. In this paper, we disclose two new relations residing in the triplets by viewing the triplet as a graph node. In analogy with the graph node, we mine two associations of text-bridged image alignment and complementary text reasoning. The text-bridged image alignment considers composed image retrieval as a specialized form of image retrieval, where the query text acts as a bridge between the query image and the target one, and a hinge-based cross attention is proposed to incorporate this relation into the network learning. On the other hand, the association of complementary text reasoning regards composed image retrieval as a specific type of cross-modal retrieval, where the composite two images are used to reason the complementary text. To integrate these views effectively, a twin attention-based compositor is designed. By combining these two types of complementary associations with the explicit query pair-target image relation, we establish a comprehensive set of constraints for composed image retrieval. With the above designs, we finally developed our CaLa, a Complementary Association Learning framework for Augmenting Composed Image Retrieval. Experimental evaluations are conducted on the widely-used CIRR and FashIionIQ benchmarks with multiple backbones to validate the effectiveness of our CaLa. The results demonstrate the superiority of our method in the composed image retrieval task. Our code and models are available at https://github.com/Chiangsonw/CaLa
- Open Access
- 24 April 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Visualization and Computer Graphics
- Vol. 30 (10) , 6984-6996
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. Supplementary videos and code can be accessed at https://zf223669.github.io/Diffmotion-v2-website/.
- 24 April 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Visualization and Computer Graphics
- Vol. 30 (10) , 6984-6996
Open AccessSpeech-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. Supplementary videos and code can be accessed at https://zf223669.github.io/Diffmotion-v2-website/.
- Open Access
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. 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. Supplementary videos and code can be accessed at https://zf223669.github.io/Diffmotion-v2-website/
Open AccessSpeech-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. 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. Supplementary videos and code can be accessed at https://zf223669.github.io/Diffmotion-v2-website/
- Open Access
- 27 February 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in TechRxiv
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. 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. Supplementary videos and code can be accessed at https://zf223669.github.io/Diffmotion-v2-website/
- 27 February 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in TechRxiv
Open AccessSpeech-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. 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. Supplementary videos and code can be accessed at https://zf223669.github.io/Diffmotion-v2-website/