A5087162987- Face recognition and analysis
- Generative Adversarial Networks and Image Synthesis
- Advanced Vision and Imaging
- 3D Shape Modeling and Analysis
- Gaze Tracking and Assistive Technology
- Facial Nerve Paralysis Treatment and Research
- Computer Graphics and Visualization Techniques
- Cultural Heritage Materials Analysis
- Video Surveillance and Tracking Methods
- Water Systems and Optimization
- Archaeological Research and Protection
- Human Pose and Action Recognition
- Image Processing and 3D Reconstruction
- Sparse and Compressive Sensing Techniques
- Advanced Image Processing Techniques
- Advanced Image and Video Retrieval Techniques
- Water Quality Monitoring Technologies
- Water resources management and optimization
- Neonatal and fetal brain pathology
Google (Switzerland)
2022-2024
ETH Zurich
2021-2024
Google (United States)
2021-2024
Dalian University of Technology
2024
University College London
2018
Duke University
2018
We propose GazeNeRF, a 3D-aware method for the task of gaze redirection. Existing redirection methods operate on 2D images and struggle to generate 3D consistent results. Instead, we build intuition that face region eyeballs are separate structures move in coordinated yet independent fashion. Our leverages recent advancements conditional image-based neural radiance fields proposes two-stream architecture predicts volumetric features eye regions separately. Rigidly transforming via rotation...
Deep generative models can synthesize photorealistic images of human faces with novel identities. However, a key challenge to the wide applicability such techniques is provide independent control over semantically meaningful parameters: appearance, head pose, face shape, and facial expressions. In this paper, we propose VariTex - best our knowledge first method that learns variational latent feature space neural textures, which allows sampling We combine model parametric gain explicit pose...
A unique challenge in creating high-quality animatable and relightable 3D avatars of people is modeling human eyes. The synthesizing eyes multifold as it requires 1) appropriate representations for the various components eye periocular region coherent viewpoint synthesis, capable representing diffuse, refractive highly reflective surfaces, 2) disentangling skin appearance from environmental illumination such that may be rendered under novel lighting conditions, 3) capturing eyeball motion...
NeRFs have enabled highly realistic synthesis of human faces including complex appearance and reflectance effects hair skin. These methods typically require a large number multi-view input images, making the process hardware intensive cumbersome, limiting applicability to unconstrained settings. We propose novel volumetric face prior that enables ultra high-resolution views subjects are not part prior's training distribution. This model consists an identity-conditioned NeRF, trained on...
Abstract Eye gaze and expressions are crucial non‐verbal signals in face‐to‐face communication. Visual effects telepresence demand significant improvements personalized tracking, animation, synthesis of the eye region to achieve true immersion. Morphable face models, combination with coordinate‐based neural volumetric representations, show promise solving difficult problem reconstructing intricate geometry (eyelashes) synthesizing photorealistic appearance variations (wrinkles specularities)...
High-fidelity, photorealistic 3D capture of a human face is long-standing problem in computer graphics – the complex material skin, intricate geometry hair, and fine scale textural details make it challenging. Traditional techniques rely on very large expensive rigs to reconstruct explicit mesh appearance maps, are limited by accuracy hand-crafted reflectance models. More recent volumetric methods (e.g., NeRFs) have enabled view-synthesis sometimes relighting learning an implicit...
ABSTRACT Frequent burst events in water distribution systems cause severe loss and other environmental issues such as contamination carbon emissions. The availability of massive monitored data has facilitated the development data-driven detection methods. This paper proposes flow subsequences clustering–reconstruction analysis method for district metering areas (DMAs). sliding window is used to create subsequence libraries all time points a day using historical set thereafter conducted...
3D rendering of dynamic face captures is a challenging problem, and it demands improvements on several fronts---photorealism, efficiency, compatibility, configurability. We present novel representation that enables high-quality volumetric an actor's facial performances with minimal compute memory footprint. It runs natively commodity graphics soft- hardware, allows for graceful trade-off between quality efficiency. Our method utilizes recent advances in neural rendering, particularly...
Human activities are inherently complex, and even simple household tasks involve numerous object interactions. To better understand these behaviors, it is crucial to model their dynamic interactions with the environment. The recent availability of affordable head-mounted cameras egocentric data offers a more accessible efficient means human-object in 3D environments. However, most existing methods for human activity modeling either focus on reconstructing models hand-object or human-scene...
Volumetric modeling and neural radiance field representations have revolutionized 3D face capture photorealistic novel view synthesis. However, these methods often require hundreds of multi-view input images are thus inapplicable to cases with less than a handful inputs. We present volumetric prior on human faces that allows for high-fidelity expressive from as few three views captured in the wild. Our key insight is an implicit trained synthetic data alone can generalize extremely...
This rapid development and testing project captured data from multiple digital imaging techniques to try see texts in papyrus mâché mummy mask cartonnage layers.Prior studies by other scholars destroyed the masks access papyri, denying future researcher primary historical artefacts.This international, multidisciplinary assessed feasibility of integrating non-destructive technologies make visible images cartonnages for open research analysis.The team used both optical multispectral coherence...
NeRFs have enabled highly realistic synthesis of human faces including complex appearance and reflectance effects hair skin. These methods typically require a large number multi-view input images, making the process hardware intensive cumbersome, limiting applicability to unconstrained settings. We propose novel volumetric face prior that enables ultra high-resolution views subjects are not part prior's training distribution. This model consists an identity-conditioned NeRF, trained on...
We propose GazeNeRF, a 3D-aware method for the task of gaze redirection. Existing redirection methods operate on 2D images and struggle to generate 3D consistent results. Instead, we build intuition that face region eyeballs are separate structures move in coordinated yet independent fashion. Our leverages recent advancements conditional image-based neural radiance fields proposes two-stream architecture predicts volumetric features eye regions separately. Rigidly transforming via rotation...
Deep generative models can synthesize photorealistic images of human faces with novel identities. However, a key challenge to the wide applicability such techniques is provide independent control over semantically meaningful parameters: appearance, head pose, face shape, and facial expressions. In this paper, we propose VariTex - best our knowledge first method that learns variational latent feature space neural textures, which allows sampling We combine model parametric gain explicit pose...