CelebV-HQ: A Large-Scale Video Facial Attributes Dataset (2022-07-25T00:00:00.000000Z)

TL;DR

This work proposes a large-scale, high-quality, and diverse video dataset with rich facial attribute annotations, named the High-Quality Celebrity Video Dataset (CelebV-HQ), and conducts a comprehensive analysis in terms of age, ethnicity, brightness stability, motion smoothness, head pose diversity, and data quality.

Abstract

Large-scale datasets have played indispensable roles in the recent success of face generation/editing and significantly facilitated the advances of emerging research fields. However, the academic community still lacks a video dataset with diverse facial attribute annotations, which is crucial for the research on face-related videos. In this work, we propose a large-scale, high-quality, and diverse video dataset with rich facial attribute annotations, named the High-Quality Celebrity Video Dataset (CelebV-HQ). CelebV-HQ contains 35,666 video clips with the resolution of 512x512 at least, involving 15,653 identities. All clips are labeled manually with 83 facial attributes, covering appearance, action, and emotion. We conduct a comprehensive analysis in terms of age, ethnicity, brightness stability, motion smoothness, head pose diversity, and data quality to demonstrate the diversity and temporal coherence of CelebV-HQ. Besides, its versatility and potential are validated on two representative tasks, i.e., unconditional video generation and video facial attribute editing. Furthermore, we envision the future potential of CelebV-HQ, as well as the new opportunities and challenges it would bring to related research directions. Data, code, and models are publicly available. Project page: https://celebv-hq.github.io.

Authors

Ziwei Liu

27 papers

Chen Change Loy

72 papers

Wayne Wu

14 papers

TL;DR

Abstract

Authors

References102 items

Expressive Talking Head Generation with Granular Audio-Visual Control

CogVideo: Large-scale Pretraining for Text-to-Video Generation via Transformers

Generalizable Neural Performer: Learning Robust Radiance Fields for Human Novel View Synthesis

Video Diffusion Models

Text2LIVE: Text-Driven Layered Image and Video Editing

TransEditor: Transformer-Based Dual-Space GAN for Highly Controllable Facial Editing

Sem2NeRF: Converting Single-View Semantic Masks to Neural Radiance Fields

Generating Videos with Dynamics-aware Implicit Generative Adversarial Networks

Stitch it in Time: GAN-Based Facial Editing of Real Videos

MobileFaceSwap: A Lightweight Framework for Video Face Swapping

StyleGAN-V: A Continuous Video Generator with the Price, Image Quality and Perks of StyleGAN2

StyleSDF: High-Resolution 3D-Consistent Image and Geometry Generation

Efficient Geometry-aware 3D Generative Adversarial Networks

HeadNeRF: A Realtime NeRF-based Parametric Head Model

NÜWA: Visual Synthesis Pre-training for Neural visUal World creAtion

Deceive D: Adaptive Pseudo Augmentation for GAN Training with Limited Data

HyperExtended LightFace: A Facial Attribute Analysis Framework

StyleNeRF: A Style-based 3D-Aware Generator for High-resolution Image Synthesis

Talk-to-Edit: Fine-Grained Facial Editing via Dialog

Alias-Free Generative Adversarial Networks

A Latent Transformer for Disentangled Face Editing in Images and Videos

Information Bottleneck Disentanglement for Identity Swapping

Animatable Neural Radiance Fields for Modeling Dynamic Human Bodies

A Good Image Generator Is What You Need for High-Resolution Video Synthesis

Pose-Controllable Talking Face Generation by Implicitly Modularized Audio-Visual Representation

Motion Representations for Articulated Animation

VideoGPT: Video Generation using VQ-VAE and Transformers

Audio-Driven Emotional Video Portraits

AD-NeRF: Audio Driven Neural Radiance Fields for Talking Head Synthesis

Designing an encoder for StyleGAN image manipulation

VisualVoice: Audio-Visual Speech Separation with Cross-Modal Consistency

FairFace: Face Attribute Dataset for Balanced Race, Gender, and Age for Bias Measurement and Mitigation

Lips Don't Lie: A Generalisable and Robust Approach to Face Forgery Detection

Dynamic Neural Radiance Fields for Monocular 4D Facial Avatar Reconstruction

pi-GAN: Periodic Implicit Generative Adversarial Networks for 3D-Aware Image Synthesis

One-Shot Free-View Neural Talking-Head Synthesis for Video Conferencing

D-NeRF: Neural Radiance Fields for Dynamic Scenes

Face Forgery Detection by 3D Decomposition

AOT: Appearance Optimal Transport Based Identity Swapping for Forgery Detection

Fast Bi-layer Neural Synthesis of One-Shot Realistic Head Avatars

TikTok

LiteFlowNet3: Resolving Correspondence Ambiguity for More Accurate Optical Flow Estimation

Emotion recognition using multi-modal data and machine learning techniques: A tutorial and review

InterFaceGAN: Interpreting the Disentangled Face Representation Learned by GANs

Facial Action Unit Intensity Estimation via Semantic Correspondence Learning with Dynamic Graph Convolution

Vision-based human action recognition: An overview and real world challenges

First Order Motion Model for Image Animation

Multi-Modal Domain Adaptation for Fine-Grained Action Recognition

Human Emotion Recognition: Review of Sensors and Methods

Deep Audio-visual Learning: A Survey

Face X-Ray for More General Face Forgery Detection

FaceShifter: Towards High Fidelity And Occlusion Aware Face Swapping

StarGAN v2: Diverse Image Synthesis for Multiple Domains

Analyzing and Improving the Image Quality of StyleGAN

Analysis of the Reasons and Development of Short Video Application——Taking Tik Tok as an Example

FSGAN: Subject Agnostic Face Swapping and Reenactment

Context-Aware Emotion Recognition Networks

Quality Assessment of In-the-Wild Videos

MaskGAN: Towards Diverse and Interactive Facial Image Manipulation

A Short Note on the Kinetics-700 Human Action Dataset

Hierarchical Cross-Modal Talking Face Generation With Dynamic Pixel-Wise Loss

Arbitrary Talking Face Generation via Attentional Audio-Visual Coherence Learning

A Style-Based Generator Architecture for Generative Adversarial Networks

Towards Accurate Generative Models of Video: A New Metric & Challenges

Dumps

Large Scale GAN Training for High Fidelity Natural Image Synthesis

ReenactGAN: Learning to Reenact Faces via Boundary Transfer

Talking Face Generation by Adversarially Disentangled Audio-Visual Representation

VoxCeleb2: Deep Speaker Recognition

Look at Boundary: A Boundary-Aware Face Alignment Algorithm

The Ryerson Audio-Visual Database of Emotional Speech and Song (RAVDESS): A dynamic, multimodal set of facial and vocal expressions in North American English

Multimodal Unsupervised Image-to-Image Translation

Looking to listen at the cocktail party

Adversarially Robust Generalization Requires More Data

FaceForensics: A Large-scale Video Dataset for Forgery Detection in Human Faces

ArcFace: Additive Angular Margin Loss for Deep Face Recognition