Multimodal Attention Fusion for Target Speaker Extraction
원문 초록
Target speaker extraction, which aims at extracting a target speaker's voice from a mixture of voices using audio, visual or locational clues, has received much interest. Recently an audio-visual target speaker extraction has been proposed that extracts target speech by using complementary audio and visual clues. Although audio-visual target speaker extraction offers a more stable performance than single modality methods for simulated data, its adaptation towards realistic situations has not been fully explored as well as evaluations on real recorded mixtures. One of the major issues to handle realistic situations is how to make the system robust to clue corruption because in real recordings both clues may not be equally reliable, e.g. visual clues may be affected by occlusions. In this work, we propose a novel attention mechanism for multi-modal fusion and its training methods that enable to effectively capture the reliability of the clues and weight the more reliable ones. Our proposals improve signal to distortion ratio (SDR) by 1.0 dB over conventional fusion mechanisms on simulated data. Moreover, we also record an audio-visual dataset of simultaneous speech with realistic visual clue corruption and show that audio-visual target speaker extraction with our proposals successfully work on real data.
함께 살펴볼 논문
공통 태그를 근거로 연결했습니다. 실제 인용 관계나 논문의 우수성을 뜻하지 않습니다.
M3ANet: Multi-scale and Multi-Modal Alignment Network for Brain-Assisted Target Speaker Extraction
LauraTSE: Target Speaker Extraction using Auto-Regressive Decoder-Only Language Models
Metis: A Foundation Speech Generation Model with Masked Generative Pre-training
Multi-Level Speaker Representation for Target Speaker Extraction
WeSep: A Scalable and Flexible Toolkit Towards Generalizable Target Speaker Extraction
Generative Speech Foundation Model Pretraining for High-Quality Speech Extraction and Restoration
토론
이 논문과 연결된 포럼 글이 아직 없습니다.