George Baciu
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
George Baciu has a total of 231 co-authors. Here are the top co-authors by number of shared publications:
Eddie C. L. Chan
23 shared publicationsJinlian Hu
Department of Computing & Institute of Textiles and Clothing, The Hong Kong Polytechnic University, Hung Hom, Kowloon, Hong Kong
13 shared publicationsChenhui Li
12 shared publicationsYu Han
Deptartment of Computing, The Hong Kong Polytechnic University, Hong Kong
10 shared publicationsYunzhe Wang
10 shared publicationsChen Xu
9 shared publications
Latest Publications
- Open Access
- 15 July 2026
- Published by Elsevier in SSRN Electronic Journal
Every Second Matters: Investigating the Effects of Conversational Response Delays and Contextualized Fillers for Digital Avatars
Emergent large language model (LLM)-based digital avatars are poised to have broad applications across domains. However, LLMs, together with auxiliary services such as text-to-speech, require processing time and can generate undesirable conversational response delays. There is still limited understanding of how such delays affect human-avatar interaction, particularly the point at which a delay changes from acceptable to unacceptable. In this paper, we first conduct a preliminary study in virtual reality (N = 30) to investigate the effects of response delays ranging from 0 to 16000 ms. The results show that delays between 0 and 2000 ms are generally acceptable, with around 1000 ms perceived as the most natural. Noticeable degradation begins by 4000 ms and becomes pronounced at 8000 ms. Based on these findings, we implemented a turn-taking scheme that uses a lightweight LLM to generate contextualized conversational fillers within 1000 ms. We then compare this approach with commonly used canned lexical and non-lexical fillers in a user study (N = 63). Canned non-lexical fillers provide no improvement and can worsen the experience, while canned lexical fillers offer only limited benefits. In contrast, contextualized fillers more effectively preserve conversational realism, social context, interactivity, anthropomorphism, and animacy, while reducing discomfort. These findings show that faster responses are not always better and that filler effectiveness depends on contextual relevance. This work provides design guidance for more natural and socially appropriate digital avatars and conversational user interfaces.
- 15 July 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessEvery Second Matters: Investigating the Effects of Conversational Response Delays and Contextualized Fillers for Digital Avatars
Emergent large language model (LLM)-based digital avatars are poised to have broad applications across domains. However, LLMs, together with auxiliary services such as text-to-speech, require processing time and can generate undesirable conversational response delays. There is still limited understanding of how such delays affect human-avatar interaction, particularly the point at which a delay changes from acceptable to unacceptable. In this paper, we first conduct a preliminary study in virtual reality (N = 30) to investigate the effects of response delays ranging from 0 to 16000 ms. The results show that delays between 0 and 2000 ms are generally acceptable, with around 1000 ms perceived as the most natural. Noticeable degradation begins by 4000 ms and becomes pronounced at 8000 ms. Based on these findings, we implemented a turn-taking scheme that uses a lightweight LLM to generate contextualized conversational fillers within 1000 ms. We then compare this approach with commonly used canned lexical and non-lexical fillers in a user study (N = 63). Canned non-lexical fillers provide no improvement and can worsen the experience, while canned lexical fillers offer only limited benefits. In contrast, contextualized fillers more effectively preserve conversational realism, social context, interactivity, anthropomorphism, and animacy, while reducing discomfort. These findings show that faster responses are not always better and that filler effectiveness depends on contextual relevance. This work provides design guidance for more natural and socially appropriate digital avatars and conversational user interfaces.
- 2 July 2026
- Published by Springer Nature
- p. 41-55
PockeTA: Virtual Teaching Assistant Platform with Course Context
- 2 July 2026
- Published by Springer Nature
- p. 41-55
PockeTA: Virtual Teaching Assistant Platform with Course Context
- 24 December 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Consumer Electronics
- Vol. 72 (1) , 192-210
A Novel Wasserstein Color Transfer Variational Framework on Fabric Images
Color transfer on fabric images has become a prevalent computer-aided tool in the fashion design and textile industry, enabling the automatic generation of simulated recolored images available for computer previews. Many existing variational methods are sourced with color palettes due to their ease to use; however, these approaches are often limited by practical inefficiencies, primarily arising from their reliance on cost-intensive iterative calculations. Moreover, the presence of color bleeding and distortion frequently degenerates the visual quality of their recolored fabric images. In this paper, we propose a novel color transfer variational framework on fabric images that consists of three modules: (1) image decomposition, (2) image segmentation, and (3) image recoloring. The first module constructs a variational Wasserstein optimization to generate independent reflectance and shading images, utilizing outputs from the intrinsic image decomposition network OS-Net for efficient modeling and initialization. Building on the piecewise constant approximation of the reflectance image, the second module formulates a variational model to partition the reflectance image into multiple regions of interest, represented by fuzzy membership functions. For image recoloring, color transfer results are produced by reintegrating the shading image with the recolored reflectance image, modified according to the given color palette. Experiment results demonstrate that, compared to classical fabric color transfer variational methods and recent learning-based approaches, our framework significantly enhances operational efficiency and achieves superior color transfer performance. It effectively eliminates unexpected color bleeding and distortion while preserving the original geometric structures and intricate fabric details in recolored images.
- 24 December 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Consumer Electronics
- Vol. 72 (1) , 192-210
A Novel Wasserstein Color Transfer Variational Framework on Fabric Images
Color transfer on fabric images has become a prevalent computer-aided tool in the fashion design and textile industry, enabling the automatic generation of simulated recolored images available for computer previews. Many existing variational methods are sourced with color palettes due to their ease to use; however, these approaches are often limited by practical inefficiencies, primarily arising from their reliance on cost-intensive iterative calculations. Moreover, the presence of color bleeding and distortion frequently degenerates the visual quality of their recolored fabric images. In this paper, we propose a novel color transfer variational framework on fabric images that consists of three modules: (1) image decomposition, (2) image segmentation, and (3) image recoloring. The first module constructs a variational Wasserstein optimization to generate independent reflectance and shading images, utilizing outputs from the intrinsic image decomposition network OS-Net for efficient modeling and initialization. Building on the piecewise constant approximation of the reflectance image, the second module formulates a variational model to partition the reflectance image into multiple regions of interest, represented by fuzzy membership functions. For image recoloring, color transfer results are produced by reintegrating the shading image with the recolored reflectance image, modified according to the given color palette. Experiment results demonstrate that, compared to classical fabric color transfer variational methods and recent learning-based approaches, our framework significantly enhances operational efficiency and achieves superior color transfer performance. It effectively eliminates unexpected color bleeding and distortion while preserving the original geometric structures and intricate fabric details in recolored images.
- Open Access
- 25 May 2025
- Published by Springer Nature in The Visual Computer
- Vol. 41 (9) , 6479-6491
Distilling complementary information from temporal context for enhancing human appearance in human-specific NeRF
Reconstructing and animating digital avatars with free views from monocular videos have been an interesting research task in the computer vision field for a long time. Recently, some methods have introduced a novel category method of leveraging the neural radiance field to represent the human body in a canonical space with the help of the SMPL model. With the deformation of the points from an observation space into a canonical space, the human appearance can be learned in various poses and viewpoints. However, previous methods highly rely on pose-dependent representation learned from frame-independent optimization and ignore the temporal contexts across the continuous motion video, causing a bad influence on the dynamic appearance texture generation. To overcome these problems, we propose a novel free-viewpoint rendering framework, TMIHuman. It aims at introducing temporal information into NeRF-based rendering and distilling task-relevant information from complex pixel-wise representations. To be specific, we build a temporal fusion encoder that imports timestamps into the learning of non-rigid deformation and fuses the visual features of other frames into human representation. Then, we propose to disentangle the fused features and extract useful visual cues via mutual information objectives. We have extensively evaluated our method and achieved state-of-the-art performance on different public datasets.
- 25 May 2025
- Published by Springer Nature in The Visual Computer
- Vol. 41 (9) , 6479-6491
Open AccessDistilling complementary information from temporal context for enhancing human appearance in human-specific NeRF
Reconstructing and animating digital avatars with free views from monocular videos have been an interesting research task in the computer vision field for a long time. Recently, some methods have introduced a novel category method of leveraging the neural radiance field to represent the human body in a canonical space with the help of the SMPL model. With the deformation of the points from an observation space into a canonical space, the human appearance can be learned in various poses and viewpoints. However, previous methods highly rely on pose-dependent representation learned from frame-independent optimization and ignore the temporal contexts across the continuous motion video, causing a bad influence on the dynamic appearance texture generation. To overcome these problems, we propose a novel free-viewpoint rendering framework, TMIHuman. It aims at introducing temporal information into NeRF-based rendering and distilling task-relevant information from complex pixel-wise representations. To be specific, we build a temporal fusion encoder that imports timestamps into the learning of non-rigid deformation and fuses the visual features of other frames into human representation. Then, we propose to disentangle the fused features and extract useful visual cues via mutual information objectives. We have extensively evaluated our method and achieved state-of-the-art performance on different public datasets.
- 24 February 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Neural Networks and Learning Systems
- Vol. 36 (8) , 14159-14173
Temporal-Interim Pose Synthesis and Distillation for Dynamic Human Pose Estimation
In the task of dynamic human pose estimation (dynamic HPE), the temporal relationships between human body parts should be captured comprehensively to understand the dynamic human motions, where the correlated motion information eventually helps to recognize body parts. The popular methods are successful in terms of utilizing long-term motion information captured by low-speed cameras. Yet they neglect the underlying intermediate motions between captured frames, which comprise the temporal-interim poses lost in the video. In this article, we introduce a novel framework, temporal-interim pose synthesis and distillation, to produce and leverage the intermediate motion information for dynamic motion establishment. The pose synthesis yields the visual feature maps of the intermediate poses, which appear between the existing video frames. It allows the synthesized and current poses to form richer motion patterns. Next, the pose distillation divides the body parts into several groups, where it learns the specific part-wise relationship within each group. It degrades the complexity of learning useful part-wise relationships from rich motion patterns and extracts more detailed motion information for fine-grained part groups. We extensively evaluate our method on challenging datasets for dynamic pose estimation, achieving state-of-the-artresults.
- 24 February 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Neural Networks and Learning Systems
- Vol. 36 (8) , 14159-14173
Temporal-Interim Pose Synthesis and Distillation for Dynamic Human Pose Estimation
In the task of dynamic human pose estimation (dynamic HPE), the temporal relationships between human body parts should be captured comprehensively to understand the dynamic human motions, where the correlated motion information eventually helps to recognize body parts. The popular methods are successful in terms of utilizing long-term motion information captured by low-speed cameras. Yet they neglect the underlying intermediate motions between captured frames, which comprise the temporal-interim poses lost in the video. In this article, we introduce a novel framework, temporal-interim pose synthesis and distillation, to produce and leverage the intermediate motion information for dynamic motion establishment. The pose synthesis yields the visual feature maps of the intermediate poses, which appear between the existing video frames. It allows the synthesized and current poses to form richer motion patterns. Next, the pose distillation divides the body parts into several groups, where it learns the specific part-wise relationship within each group. It degrades the complexity of learning useful part-wise relationships from rich motion patterns and extracts more detailed motion information for fine-grained part groups. We extensively evaluate our method on challenging datasets for dynamic pose estimation, achieving state-of-the-artresults.
- 23 December 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Multimedia
- Vol. 27 (15209210) , 1706-1721
Accurate-PGNet: Learning to Assemble Perceptual Body Parts for Accurate Human Skeleton Establishment
The human skeleton establishment aims to provide accurate localization information of the human body from RGB images and establish a complete human skeleton for many applications, such as action recognition, video surveillance, and human-computer interaction. Considering the inherent human body structure, many recent methods group the relevant body parts and utilize the deep convolutional network to learn the visual context from the part groups. However, the grouping approaches used in these methods heavily rely on prior knowledge of the human body shape but lose important relationships between parts. In this paper, we introduce the Accurate Part Grouping Network (Accurate-PGNet), a novel network for hierarchically grouping body parts in a data-driven manner. In contrast to the previous methods, we use neural architecture search (NAS) to optimize the architecture of Accurate-PGNet and properly group the body parts. The part grouping respects the diverse visual patterns of parts, producing groups containing different body parts. From each group, we learn the visual feature map. It helps to capture the correlation between parts and predict their locations. The feature maps of the part groups are merged hierarchically to capture the higher-order context of parts in larger groups. We extensively evaluated our method on the challenging benchmarks, demonstrating that Accurate-PGNet effectively helps to achieve state-of-the-art results.
- 23 December 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Multimedia
- Vol. 27 (15209210) , 1706-1721
Accurate-PGNet: Learning to Assemble Perceptual Body Parts for Accurate Human Skeleton Establishment
The human skeleton establishment aims to provide accurate localization information of the human body from RGB images and establish a complete human skeleton for many applications, such as action recognition, video surveillance, and human-computer interaction. Considering the inherent human body structure, many recent methods group the relevant body parts and utilize the deep convolutional network to learn the visual context from the part groups. However, the grouping approaches used in these methods heavily rely on prior knowledge of the human body shape but lose important relationships between parts. In this paper, we introduce the Accurate Part Grouping Network (Accurate-PGNet), a novel network for hierarchically grouping body parts in a data-driven manner. In contrast to the previous methods, we use neural architecture search (NAS) to optimize the architecture of Accurate-PGNet and properly group the body parts. The part grouping respects the diverse visual patterns of parts, producing groups containing different body parts. From each group, we learn the visual feature map. It helps to capture the correlation between parts and predict their locations. The feature maps of the part groups are merged hierarchically to capture the higher-order context of parts in larger groups. We extensively evaluated our method on the challenging benchmarks, demonstrating that Accurate-PGNet effectively helps to achieve state-of-the-art results.
- 27 August 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Neural Networks and Learning Systems
- Vol. 36 (5) , 9195-9209
DSANet: Dynamic and Structure-Aware GCN for Sparse and Incomplete Point Cloud Learning
Learning 3-D structures from incomplete point clouds with extreme sparsity and random distributions is a challenge since it is difficult to infer topological connectivity and structural details from fragmentary representations. Missing large portions of informative structures further aggravates this problem. To overcome this, a novel graph convolutional network (GCN) called dynamic and structure-aware NETwork (DSANet) is presented in this article. This framework is formulated based on a pyramidic auto-encoder (AE) architecture to address accurate structure reconstruction on the sparse and incomplete point clouds. A PointNet-like neural network is applied as the encoder to efficiently aggregate the global representations of coarse point clouds. On the decoder side, we design a dynamic graph learning module with a structure-aware attention (SAA) to take advantage of the topology relationships maintained in the dynamic latent graph. Relying on gradually unfolding the extracted representation into a sequence of graphs, DSANet is able to reconstruct complicated point clouds with rich and descriptive details. To associate analogous structure awareness with semantic estimation, we further propose a mechanism, called structure similarity assessment (SSA). This method allows our model to surmise semantic homogeneity in an unsupervised manner. Finally, we optimize the proposed model by minimizing a new distortion-aware objective end-to-end. Extensive qualitative and quantitative experiments demonstrate the impressive performance of our model in reconstructing unbroken 3-D shapes from deficient point clouds and preserving semantic relationships among different regional structures.
- 27 August 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Neural Networks and Learning Systems
- Vol. 36 (5) , 9195-9209
DSANet: Dynamic and Structure-Aware GCN for Sparse and Incomplete Point Cloud Learning
Learning 3-D structures from incomplete point clouds with extreme sparsity and random distributions is a challenge since it is difficult to infer topological connectivity and structural details from fragmentary representations. Missing large portions of informative structures further aggravates this problem. To overcome this, a novel graph convolutional network (GCN) called dynamic and structure-aware NETwork (DSANet) is presented in this article. This framework is formulated based on a pyramidic auto-encoder (AE) architecture to address accurate structure reconstruction on the sparse and incomplete point clouds. A PointNet-like neural network is applied as the encoder to efficiently aggregate the global representations of coarse point clouds. On the decoder side, we design a dynamic graph learning module with a structure-aware attention (SAA) to take advantage of the topology relationships maintained in the dynamic latent graph. Relying on gradually unfolding the extracted representation into a sequence of graphs, DSANet is able to reconstruct complicated point clouds with rich and descriptive details. To associate analogous structure awareness with semantic estimation, we further propose a mechanism, called structure similarity assessment (SSA). This method allows our model to surmise semantic homogeneity in an unsupervised manner. Finally, we optimize the proposed model by minimizing a new distortion-aware objective end-to-end. Extensive qualitative and quantitative experiments demonstrate the impressive performance of our model in reconstructing unbroken 3-D shapes from deficient point clouds and preserving semantic relationships among different regional structures.
- 8 June 2024
- Published by Springer Nature in The Visual Computer
- Vol. 41 (3) , 1907-1920
Fabric image recolorization by fuzzy pretrained neural network
- 8 June 2024
- Published by Springer Nature in The Visual Computer
- Vol. 41 (3) , 1907-1920
Fabric image recolorization by fuzzy pretrained neural network
- Open Access
- 28 May 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Learning Technologies
- Vol. 17, 1794-1808
Knowledge-Graph-Driven Mind Mapping for Immersive Collaborative Learning: A Pilot Study in Edu-Metaverse
One of the promises of edu-metaverse is its ability to provide a virtual environment that enables us to engage in learning activities that are similar to or on par with reality. The digital enhancements introduced in a virtual environment contribute to our increased expectations of novel learning experiences. However, despite its promising outcomes, there appears to be limited adoption of the edu-metaverse for practical learning at this time. We believe this can be attributed to the fact that there is a lack of investigation into learners' behavior given a social learning environment. This lack of investigation is critical, as without behavioral insight, it hinders the development of education material and the direction of an edu-metaverse. This paper aims to provide insights into learner behavior in edu-metaverse by conducting a focused pilot study. Particularly, we investigate the learners' behavior when drawing mind maps in the metaverse aided by the embedded knowledge graph. Compared to Zoom, a typical video conferencing and remote collaboration platform, it seems that users are more involved in learning activities, and the co-presence seems to assist in engagement so that they are more willing to contribute to the learning task collaboratively. Our work also provides insight into the interaction and learning activity design for edu-metaverse, particularly when mind maps are involved.
- 28 May 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Learning Technologies
- Vol. 17, 1794-1808
Open AccessKnowledge-Graph-Driven Mind Mapping for Immersive Collaborative Learning: A Pilot Study in Edu-Metaverse
One of the promises of edu-metaverse is its ability to provide a virtual environment that enables us to engage in learning activities that are similar to or on par with reality. The digital enhancements introduced in a virtual environment contribute to our increased expectations of novel learning experiences. However, despite its promising outcomes, there appears to be limited adoption of the edu-metaverse for practical learning at this time. We believe this can be attributed to the fact that there is a lack of investigation into learners' behavior given a social learning environment. This lack of investigation is critical, as without behavioral insight, it hinders the development of education material and the direction of an edu-metaverse. This paper aims to provide insights into learner behavior in edu-metaverse by conducting a focused pilot study. Particularly, we investigate the learners' behavior when drawing mind maps in the metaverse aided by the embedded knowledge graph. Compared to Zoom, a typical video conferencing and remote collaboration platform, it seems that users are more involved in learning activities, and the co-presence seems to assist in engagement so that they are more willing to contribute to the learning task collaboratively. Our work also provides insight into the interaction and learning activity design for edu-metaverse, particularly when mind maps are involved.
- 24 November 2023
- Published by Springer Nature
- p. 125-138
Gradual Study Advising with Course Knowledge Graphs
- 24 November 2023
- Published by Springer Nature
- p. 125-138
Gradual Study Advising with Course Knowledge Graphs