| ROAM-ASD: Robust Open-World Active Speaker Detection with Flexible Multimodal Fusion | | 96.5% | ROAM-ASD | 2026-09-22 |
| TalkNCE: Improving Active Speaker Detection with Talk-Aware Contrastive Learning | ✓ Link | 95.5% | LoCoNet+TalkNCE | 2023-09-21 |
| LASER: Lip Landmark Assisted Speaker Detection for Robustness | ✓ Link | 95.3% | LoCoNet + Laser | 2025-01-21 |
| LoCoNet: Long-Short Context Network for Active Speaker Detection | ✓ Link | 95.2% | LoCoNet | 2023-01-19 |
| GateFusion: Hierarchical Gated Cross-Modal Fusion for Active Speaker Detection | | 95.0% | GateFusion | 2025-12-17 |
| Learning Long-Term Spatial-Temporal Graphs for Active Speaker Detection | ✓ Link | 94.9% | SPELL+ | 2022-07-15 |
| UniCon+: ICTCAS-UCAS Submission to the AVA-ActiveSpeaker Task at ActivityNet Challenge 2022 | | 94.5% | UniCon+ | 2022-06-22 |
| Learning Long-Term Spatial-Temporal Graphs for Active Speaker Detection | ✓ Link | 94.2% | SPELL | 2022-07-15 |
| End-to-End Active Speaker Detection | ✓ Link | 94.1% | EASEE-50 | 2022-03-27 |
| A Light Weight Model for Active Speaker Detection | ✓ Link | 94.1% | Light-ASD | 2023-03-08 |
| ICTCAS-UCAS-TAL Submission to the AVA-ActiveSpeaker Task at ActivityNet Challenge 2021 | | 93.6% | Extended UniCon | 2021-06-01 |
| How to Design a Three-Stage Architecture for Audio-Visual Active Speaker Detection in the Wild | ✓ Link | 93.5% | ASDNet | 2021-06-07 |
| Audio-Visual Activity Guided Cross-Modal Identity Association for Active Speaker Detection | ✓ Link | 92.86% | GSCMIA | 2022-12-01 |
| NUS-HLT Report for ActivityNet Challenge 2021 AVA (Speaker) | ✓ Link | 92.3% | TalkNet | 2021-06-01 |
| UniCon: Unified Context Network for Robust Active Speaker Detection | | 92.0% | UniCon | 2021-08-05 |
| Active Speaker Detection as a Multi-Objective Optimization with Uncertainty-based Multimodal Fusion | | 91.9% | SA-uncertainty Fusion | 2021-06-07 |
| Sub-word Level Lip Reading With Visual Attention | | 89.2% | VTP (visual only) | 2021-10-14 |
| MAAS: Multi-modal Assignation for Active Speaker Detection | ✓ Link | 88.8% | MAAS-TAN | 2021-01-11 |
| Naver at ActivityNet Challenge 2019 -- Task B Active Speaker Detection (AVA) | | 87.8% | VGG-{LSTM+TCN} (ensemble) | 2019-06-25 |
| Active Speakers in Context | ✓ Link | 87.1% | Active Speakers in Context | 2020-05-20 |
| MAAS: Multi-modal Assignation for Active Speaker Detection | ✓ Link | 85.1% | MAAS-LAN | 2021-01-11 |
| Multi-Task Learning for Audio Visual Active Speaker Detection | | 84.0% | 3D-ResNet-GRU | 2019-06-01 |