A5060470295- Advanced Vision and Imaging
- 3D Shape Modeling and Analysis
- Computer Graphics and Visualization Techniques
- Human Pose and Action Recognition
- Generative Adversarial Networks and Image Synthesis
- Industrial Vision Systems and Defect Detection
- Advanced Image Processing Techniques
- Human Motion and Animation
- Structural Health Monitoring Techniques
- Welding Techniques and Residual Stresses
- Image and Object Detection Techniques
- Interactive and Immersive Displays
- Speech and Audio Processing
- Model-Driven Software Engineering Techniques
- RFID technology advancements
- Millimeter-Wave Propagation and Modeling
- Non-Destructive Testing Techniques
- Power Line Communications and Noise
- Metallurgy and Material Forming
- Domain Adaptation and Few-Shot Learning
- 3D Surveying and Cultural Heritage
- Gear and Bearing Dynamics Analysis
- Mechanical Engineering and Vibrations Research
- Direction-of-Arrival Estimation Techniques
- Wireless Communication Networks Research
Chongqing University
2022-2023
Fudan University
2011-2022
Xidian University
2017
Northwest Institute of Nuclear Technology
2007
We study the problem of shape generation in 3D mesh representation from a few color images with known camera poses. While many previous works learn to hallucinate directly priors, we resort further improving quality by leveraging cross-view information graph convolutional network. Instead building direct mapping function shape, our model learns predict series deformations improve coarse iteratively. Inspired traditional multiple view geometry methods, network samples nearby area around...
Pose transfer has been studied for decades, in which the pose of a source mesh is applied to target mesh. Particularly this paper, we are interested transferring human deform mesh, while and meshes may have different identity information. Traditional studies assume that paired existed with point-wise correspondences user annotated landmarks/mesh points, requires heavy labelling efforts. On other hand, generalization ability deep models limited, when identities. To break limitation, proposes...
We study the problem of shape generation in 3D mesh representation from a small number color images with or without camera poses. While many previous works learn to hallucinate directly priors, we adopt further improve quality by leveraging cross-view information graph convolution network. Instead building direct mapping function shape, our model learns predict series deformations coarse iteratively. Inspired traditional multiple view geometry methods, network samples nearby area around...
Generating realistic images with the guidance of reference and human poses is challenging. Despite success previous works on synthesizing person in iconic views, no efforts are made towards task pose-guided image synthesis non-iconic views. Particularly, we find that models cannot handle such a complex task, where captured views by commercially-available digital cameras. To this end, propose new framework - Multi-branch Refinement Network (MR-Net), which utilizes several visual cues,...
In this paper, we present a novel 3D head avatar creation approach capable of generalizing from few-shot in-the-wild data with high-fidelity and animatable robustness. Given the underconstrained nature problem, incorporating prior knowledge is essential. Therefore, propose framework comprising learning phases. The phase leverages priors derived large-scale multi-view dynamic dataset, applies these for personalization. Our effectively captures by utilizing Gaussian Splatting-based...
Pose transfer has been studied for decades, in which the pose of a source mesh is applied to target mesh. Particularly this paper, we are interested transferring human deform mesh, while and meshes may have different identity information. Traditional studies assume that paired existed with point-wise correspondences user annotated landmarks/mesh points, requires heavy labelling efforts. On other hand, generalization ability deep models limited, when identities. To break limitation, proposes...