Looking into Your Speech: Learning Cross-modal Affinity for Audio-visual Speech Separation
In this paper, we address the problem of separating individual speech signals from videos using audio-visual neural processing. Most conventional approaches utilize frame-wise matching criteria to extract shared information between audio and video signals; thus, their performance heavily depends on the accuracy of audio-visual synchronization and the effectiveness of their representations. To overcome the frame discontinuity problem between two modalities due to transmission delay mismatch or jitter, we propose a cross-modal affinity network (CaffNet) that learns global correspondence as well as locally-varying affinities between audio and visual streams. Since the global term provides stability over a temporal sequence at the utterance-level, this also resolves a label permutation problem characterized by inconsistent assignments. By introducing a complex convolution network, CaffNet-C, that estimates both magnitude and phase representations in the time-frequency domain, we further improve the separation performance. Experimental results verify that the proposed methods outperform conventional ones on various datasets, demonstrating their advantages in real-world scenarios.
*Jiyoung Lee and Soo-Whan Chung contributed equally to this work.
**Hong-Goo Kang and Kwanghoon Sohn are co-corresponding authors.
|326||International Conference||Hyeon-Kyeong Shin, Hyewon Han, Doyeon Kim, Soo-Whan Chung, Hong-Goo Kang "Learning Audio-Text Agreement for Open-vocabulary Keyword Spotting" in INTERSPEECH, 2022|
|325||International Conference||Changhwan Kim, Se-yun Um, Hyungchan Yoon, Hong-goo Kang "FluentTTS: Text-dependent Fine-grained Style Control for Multi-style TTS" in INTERSPEECH, 2022|
|324||International Conference||Miseul Kim, Zhenyu Piao, Seyun Um, Ran Lee, Jaemin Joh, Seungshin Lee, Hong-Goo Kang "Light-Weight Speaker Verification with Global Context Information" in INTERSPEECH, 2022|
|323||International Journal||Kyungguen Byun, Se-yun Um, Hong-Goo Kang "Length-Normalized Representation Learning for Speech Signals" in IEEE Access, vol.10, pp.60362-60372, 2022|
|322||International Conference||Doyeon Kim, Hyewon Han, Hyeon-Kyeong Shin, Soo-Whan Chung, Hong-Goo Kang "Phase Continuity: Learning Derivatives of Phase Spectrum for Speech Enhancement" in ICASSP, 2022|
|321||International Conference||Chanwoo Lee, Hyungseob Lim, Jihyun Lee, Inseon Jang, Hong-Goo Kang "Progressive Multi-Stage Neural Audio Coding with Guided References" in ICASSP, 2022|
|320||International Conference||Jihyun Lee, Hyungseob Lim, Chanwoo Lee, Inseon Jang, Hong-Goo Kang "Adversarial Audio Synthesis Using a Harmonic-Percussive Discriminator" in ICASSP, 2022|
|319||International Conference||Jinyoung Lee and Hong-Goo Kang "Stacked U-Net with High-level Feature Transfer for Parameter Efficient Speech Enhancement" in APSIPA ASC, 2021|
|318||International Conference||Huu-Kim Nguyen, Kihyuk Jeong, Se-Yun Um, Min-Jae Hwang, Eunwoo Song, Hong-Goo Kang "LiteTTS: A Decoder-free Light-weight Text-to-wave Synthesis Based on Generative Adversarial Networks" in INTERSPEECH, 2021|
|317||International Conference||Zainab Alhakeem, Yoohwan Kwon, Hong-Goo Kang "Disentangled Representations for Arabic Dialect Identification based on Supervised Clustering with Triplet Loss" in EUSIPCO, 2021|