AI Paper Atlas

Research map / Benchmarks / HDTF

HDTF leaderboard

46 papers at ICML, NeurIPS, ICLR, CVPR, AAAI, ICCV and ECCV, and the journals TPAMI, IJCV and JMLR report results on HDTF. Best FID: 7.031 (X-Dub G_free, ICML 2026).

Numbers are as each paper reports them, read by AI from its results tables and not checked by hand. Papers can differ in setting (training data, model size, extra data); the topic column shows what each paper is about. Open the paper before you compare or quote a number.

HDTF · FID

36 papers, each with its best reported number for its own method; lower is better.

26 more papers on this leaderboard, the baselines reported on it and a CSV download are free with a scime account.
Sign in free to see all

HDTF · FVD

20 papers, each with its best reported number for its own method; lower is better.

#FVD ↓MethodPaperTopic
136.93VoCa (Wan2.2-5B)VoCa: Unified Autoregressive Modeling for Talking Audio-Video Generation
ECCV 2026 · Table 1
Video generation
267.66EMO
Stable Diffusion 1.5
EMO: Emote Portrait Alive - Generating Expressive Portrait Videos with Audio2Video Diffusion Model under Weak Conditions
ECCV 2024 · Table 1
Video generation
369.04OursIM-Portrait: Learning 3D-aware Video Diffusion for Photorealistic Talking Heads from Monocular VideosC
CVPR 2025 · Table 1
3D generation
481.86Ours (w/ DPO)
Wan2.1-I2V-14B
Training-Free Multi-Character Audio-Driven Animation via Diffusion Transformer with Reward Feedback
AAAI 2026 · Table 1
Video generation
5121.8Ours
Hallo
Lightning Fast Caching-based Parallel Denoising Prediction for Accelerating Talking Head Generation
AAAI 2026 · Table 1
Video generation
6137.25KeyFaceKeyFace: Expressive Audio-Driven Facial Animation for Long Sequences via KeyFrame Interpolation
CVPR 2025 · Table 1
Video generation
7160.838Hallo3
CogVideoX
Hallo3: Highly Dynamic and Realistic Portrait Image Animation with Video Diffusion Transformer
CVPR 2025 · Table 1
Video generation
8173.463TellerTeller: Real-Time Streaming Audio-Driven Portrait Animation with Autoregressive Motion Generation
CVPR 2025 · Table 1
Video generation
9176.63X-Dub G_freeFrom Inpainting to Editing: Unlocking Robust Mask-Free Visual Dubbing via Generative Bootstrapping
ICML 2026 · Table 1
Video generation
10178.73InterTalkTowards Flexible, Natural, Efficient Interaction for Conversational Talking Face Generation
ECCV 2026 · Table 1
Video generation
10 more papers on this leaderboard, the baselines reported on it and a CSV download are free with a scime account.
Sign in free to see all

HDTF · Sync-C

17 papers, each with its best reported number for its own method; higher is better.

#Sync-C ↑MethodPaperTopic
18.658OursREAD: Real-time and Efficient Asynchronous Diffusion for Audio-driven Talking Head Generation
AAAI 2026 · Table 1
Video generation
28.59InfinityHuman
Goku-I2V
InfinityHuman: Towards Long-Term Audio-Driven Human Animation
CVPR 2026 · Table 1
Video generation
38.58PlaymatePlaymate: Flexible Control of Portrait Animation via 3D-Implicit Space Guided Diffusion
ICML 2025 · Table 1
Video generation
48.562X-Dub G_freeFrom Inpainting to Editing: Unlocking Robust Mask-Free Visual Dubbing via Generative Bootstrapping
ICML 2026 · Table 1
Video generation
58.54MultiTalk-single
Wan2.1-I2V-14B
Let Them Talk: Audio-Driven Multi-Person Conversational Video Generation
NeurIPS 2025 · Table: talking head comparison
Video generation
68.335REST (Ours)REST: Diffusion-based Real-time End-to-end Streaming Talking Head Generation via ID-Context Caching and Asynchronous Streaming Distillation
ICML 2026 · Table 1
Video generation
78.3InterTalkTowards Flexible, Natural, Efficient Interaction for Conversational Talking Face Generation
ECCV 2026 · Table 1
Video generation
88.15Ours (w/ DPO)
Wan2.1-I2V-14B
Training-Free Multi-Character Audio-Driven Animation via Diffusion Transformer with Reward Feedback
AAAI 2026 · Table 1
Video generation
98.054VoCa (Wan2.2-5B)VoCa: Unified Autoregressive Modeling for Talking Audio-Video Generation
ECCV 2026 · Table 1
Video generation
108.04SyncDreamer
Wan DiT
SyncDreamer: Controllable and Expressive Avatar Generation Beyond the Talking Head
CVPR 2026 · Table 1
Video generation
7 more papers on this leaderboard, the baselines reported on it and a CSV download are free with a scime account.
Sign in free to see all

HDTF · LPIPS

13 papers, each with its best reported number for its own method; lower is better.

3 more papers on this leaderboard, the baselines reported on it and a CSV download are free with a scime account.
Sign in free to see all

HDTF · Sync-D

12 papers, each with its best reported number for its own method; lower is better.

#Sync-D ↓MethodPaperTopic
16.15SyncDreamer
Wan DiT
SyncDreamer: Controllable and Expressive Avatar Generation Beyond the Talking Head
CVPR 2026 · Table 1
Video generation
26.69MultiTalk-single
Wan2.1-I2V-14B
Let Them Talk: Audio-Driven Multi-Person Conversational Video Generation
NeurIPS 2025 · Table: talking head comparison
Video generation
36.89OursREAD: Real-time and Efficient Asynchronous Diffusion for Audio-driven Talking Head Generation
AAAI 2026 · Table 1
Video generation
46.91InterTalkTowards Flexible, Natural, Efficient Interaction for Conversational Talking Face Generation
ECCV 2026 · Table 1
Video generation
56.985PlaymatePlaymate: Flexible Control of Portrait Animation via 3D-Implicit Space Guided Diffusion
ICML 2025 · Table 1
Video generation
67.32Ours (w/ DPO)
Wan2.1-I2V-14B
Training-Free Multi-Character Audio-Driven Animation via Diffusion Transformer with Reward Feedback
AAAI 2026 · Table 1
Video generation
77.536TellerTeller: Real-Time Streaming Audio-Driven Portrait Animation with Autoregressive Motion Generation
CVPR 2025 · Table 1
Video generation
87.697Ours
Hallo
Hallo2: Long-Duration and High-Resolution Audio-Driven Portrait Image Animation
ICLR 2025 · text
Video generation
97.925Ours-Balanced
Stable Diffusion 1.5
FADA: Fast Diffusion Avatar Synthesis with Mixed-Supervised Multi-CFG Distillation
CVPR 2025 · Table 1
Fast diffusion sampling
108.1026MegActor-ΣMegActor-Sigma: Unlocking Flexible Mixed-Modal Control in Portrait Animation with Diffusion Transformer
AAAI 2025 · Table: quantitative comparison
Cross-modal learning
2 more papers on this leaderboard, the baselines reported on it and a CSV download are free with a scime account.
Sign in free to see all

HDTF · SSIM

12 papers, each with its best reported number for its own method; higher is better.

2 more papers on this leaderboard, the baselines reported on it and a CSV download are free with a scime account.
Sign in free to see all

Not affiliated with any conference, journal or the benchmark's authors. Corrections: contact@latenticlabs.com.