Shared Multi-modal Embedding Space for Face-Voice Association
Not on the map. Why?
Abstract (arXiv)
The FAME 2026 challenge comprises two demanding tasks: training face-voice associations combined with a multilingual setting that includes testing on languages on which the model was not trained. Our approach consists of separate uni-modal processing pipelines with general face and voice feature extraction, complemented by additional age-gender feature extraction to support prediction. The resulting single-modal features are projected into a shared embedding space and trained with an Adaptive Angular Margin (AAM) loss. Our approach achieved first place in the FAME 2026 challenge, with an average Equal-Error Rate (EER) of 23.99%.
Key facts. Cited by 2 papers in the map (data through 7 Oct 2026). Builds on 8 of its 14 references that are in the map. First posted on arXiv 4 Dec 2025.
Builds on
8 of 14 references.
References · 8
- 2015FaceNet: A Unified Embedding for Face Recognition and Clusteringcited by 2,833
- 2017VGGFace2: A dataset for recognising faces across pose and agecited by 911
- 2018VoxCeleb2: Deep Speaker Recognitioncited by 862
- 2020ECAPA-TDNN: Emphasized Channel Attention, Propagation and Aggregation in TDNN Based Speaker Verificationcited by 606
- 2021Fusion and Orthogonal Projection for Improved Face-Voice Associationcited by 29
- 2022Speaker recognition with two-step multi-modal deep cleansingcited by 6
- 2025Face-voice Association in Multilingual Environments (FAME) 2026 Challenge Evaluation Plancited by 6
- 2024Multi-Stage Face-Voice Association Learning with Keynote Speaker Diarizationcited by 3
Built on by
Citing papers per year: 2025: 1; 2026: 1.
2 papers.