论文检索

Minsu Kim, Jeong Hun Yeo, Jeongsoo Choi, Yong Man Ro

This paper proposes a novel lip reading framework, especially for low-resource languages, which has not been well addressed in the previous literature. Since low-resource languages do not have enough video-text paired data to train the model to have sufficient power to model lip movements and language, it is regarded as challenging to develop lip reading models for low-resource languages. In order to mitigate the challenge, we try to learn general speech knowledge, the ability to model lip movements, from a high-resource language through the prediction of speech units. It is known that different languages partially share common phonemes, thus general speech knowledge learned from one language can be extended to other languages. Then, we try to learn language-specific knowledge, the ability to model language, by proposing Language-specific Memory-augmented Decoder (LMDecoder). LMDecoder saves language-specific audio features into memory banks and can be trained on audio-text paired data which is more easily accessible than video-text paired data. Therefore, with LMDecoder, we can transform the input speech units into language-specific audio features and translate them into texts by utilizing the learned rich language knowledge. Finally, by combining general speech knowledge and language-specific knowledge, we can efficiently develop lip reading models even for low-resource languages. Through extensive experiments using five languages, English, Spanish, French, Italian, and Portuguese, the effectiveness of the proposed method is evaluated.

Urwa Muaz, Wondong Jang, Rohun Tripathi, Santhosh Mani, Wenbin Ouyang, Ravi Teja Gadde, Baris Gecer, Sergio Elizondo, Reza Madad, Naveen Nair

Dubbed video generation aims to accurately synchronize mouth movements of a given facial video with driving audio while preserving identity and scene-specific visual dynamics, such as head pose and lighting. Despite the accurate lip generation of previous approaches that adopts a pretrained audio-video synchronization metric as an objective function, called Sync-Loss, extending it to high-resolution videos was challenging due to shift biases in the loss landscape that inhibit tandem optimization of Sync-Loss and visual quality, leading to a loss of detail. To address this issue, we introduce shift-invariant learning, which generates photo-realistic high-resolution videos with accurate Lip-Sync. Further, we employ a pyramid network with coarse-to-fine image generation to improve stability and lip syncronization. Our model outperforms state-of-the-art methods on multiple benchmark datasets, including AVSpeech, HDTF, and LRW, in terms of photo-realism, identity preservation, and Lip-Sync accuracy.

Tengda Han, Max Bain, Arsha Nagrani, Gul Varol, Weidi Xie, Andrew Zisserman

Audio Description (AD) is the task of generating descriptions of visual content, at suitable time intervals, for the benefit of visually impaired audiences. For movies, this presents notable challenges -- AD must occur only during existing pauses in dialogue, should refer to characters by name, and ought to aid understanding of the storyline as a whole. To this end, we develop a new model for automatically generating movie AD, given CLIP visual features of the frames, the cast list, and the temporal locations of the speech; addressing all three of the `who', `when', and `what' questions: (i) who -- we introduce a character bank consisting of the character's name, the actor that played the part, and a CLIP feature of their face, for the principal cast of each movie, and demonstrate how this can be used to improve naming in the generated AD; (ii) when -- we investigate several models for determining whether an AD should be generated for a time interval or not, based on the visual content of the interval and its neighbours; and (iii) what -- we implement a new vision-language model for this task, that can ingest the proposals from the character bank, whilst conditioning on the visual features using cross-attention, and demonstrate how this improves over previous architectures for AD text generation in an apples-to-apples comparison.

Jeongsoo Choi, Joanna Hong, Yong Man Ro

Recent research has demonstrated impressive results in video-to-speech synthesis which involves reconstructing speech solely from visual input. However, previous works have struggled to accurately synthesize speech due to a lack of sufficient guidance for the model to infer the correct content with the appropriate sound. To resolve the issue, they have adopted an extra speaker embedding as a speaking style guidance from a reference auditory information. Nevertheless, it is not always possible to obtain the audio information from the corresponding video input, especially during the inference time. In this paper, we present a novel vision-guided speaker embedding extractor using a self-supervised pre-trained model and prompt tuning technique. In doing so, the rich speaker embedding information can be produced solely from input visual information, and the extra audio information is not necessary during the inference time. Using the extracted vision-guided speaker embedding representations, we further develop a diffusion-based video-to-speech synthesis model, so called DiffV2S, conditioned on those speaker embeddings and the visual representation extracted from the input video. The proposed DiffV2S not only maintains phoneme details contained in the input video frames, but also creates a highly intelligible mel-spectrogram in which the speaker identities of the multiple speakers are all preserved. Our experimental results show that DiffV2S achieves the state-of-the-art performance compared to the previous video-to-speech synthesis technique.

Jiahe Li, Jiawei Zhang, Xiao Bai, Jun Zhou, Lin Gu

This paper presents ER-NeRF, a novel conditional Neural Radiance Fields (NeRF) based architecture for talking portrait synthesis that can concurrently achieve fast convergence, real-time rendering, and state-of-the-art performance with small model size. Our idea is to explicitly exploit the unequal contribution of spatial regions to guide talking portrait modeling. Specifically, to improve the accuracy of dynamic head reconstruction, a compact and expressive NeRF-based Tri-Plane Hash Representation is introduced by pruning empty spatial regions with three planar hash encoders. For speech audio, we propose a Region Attention Module to generate region-aware condition feature via an attention mechanism. Different from existing methods that utilize an MLP-based encoder to learn the cross-modal relation implicitly, the attention mechanism builds an explicit connection between audio features and spatial regions to capture the priors of local motions. Moreover, a direct and fast Adaptive Pose Encoding is introduced to optimize the head-torso separation problem by mapping the complex transformation of the head pose into spatial coordinates. Extensive experiments demonstrate that our method renders better high-fidelity and audio-lips synchronized talking portrait videos, with realistic details and high efficiency compared to previous methods.

Xinlin Li, Bang Liu, Rui Heng Yang, Vanessa Courville, Chao Xing, Vahid Partovi Nia

Efficiently deploying deep neural networks on low-resource edge devices is challenging due to their ever-increasing resource requirements. To address this issue, researchers have proposed multiplication-free neural networks, such as Power-of-Two quantization, or also known as Shift networks, which aim to reduce memory usage and simplify computation. However, existing low-bit Shift networks are not as accurate as their full-precision counterparts, typically suffering from limited weight range encoding schemes and quantization loss. In this paper, we propose the DenseShift network, which significantly improves the accuracy of Shift networks, achieving competitive performance to full-precision networks for vision and speech applications. In addition, we introduce a method to deploy an efficient DenseShift network using non-quantized floating-point activations, while obtaining 1.6X speed-up over existing methods. To achieve this, we demonstrate that zero-weight values in low-bit Shift networks do not contribute to model capacity and negatively impact inference computation. To address this issue, we propose a zero-free shifting mechanism that simplifies inference and increases model capacity. We further propose a sign-scale decomposition design to enhance training efficiency and a low-variance random initialization strategy to improve the model's transfer learning performance. Our extensive experiments on various computer vision and speech tasks demonstrate that DenseShift outperforms existing low-bit multiplication-free networks and achieves competitive performance compared to full-precision networks. Furthermore, our proposed approach exhibits strong transfer learning performance without a drop in accuracy. Our code was released on GitHub.

Ziqiao Peng, Haoyu Wu, Zhenbo Song, Hao Xu, Xiangyu Zhu, Jun He, Hongyan Liu, Zhaoxin Fan

Speech-driven 3D face animation aims to generate realistic facial expressions that match the speech content and emotion. However, existing methods often neglect emotional facial expressions or fail to disentangle them from speech content. To address this issue, this paper proposes an end-to-end neural network to disentangle different emotions in speech so as to generate rich 3D facial expressions. Specifically, we introduce the emotion disentangling encoder (EDE) to disentangle the emotion and content in the speech by cross-reconstructed speech signals with different emotion labels. Then an emotion-guided feature fusion decoder is employed to generate a 3D talking face with enhanced emotion. The decoder is driven by the disentangled identity, emotional, and content embeddings so as to generate controllable personal and emotional styles. Finally, considering the scarcity of the 3D emotional talking face data, we resort to the supervision of facial blendshapes, which enables the reconstruction of plausible 3D faces from 2D emotional data, and contribute a large-scale 3D emotional talking face dataset (3D-ETF) to train the network. Our experiments and user studies demonstrate that our approach outperforms state-of-the-art methods and exhibits more diverse facial movements. We recommend watching the supplementary video: https://ziqiaopeng. github.io/emotalk

Zheng Wang 0046, Cheng Long 0001, Shihao Xu, Bingzheng Gan, Wei Shi, Zhao Cao, Tat-Seng Chua

A large language model is a type of artificial intelligence model designed to understand and generate natural language text, such as GPT, T5, RoBERTa, BERT, etc. These models are trained on vast amounts of text data, allowing them to learn the patterns and structures of human language. With the increasing amount of multimodal information such as audio, visual, and text data generated, there is a growing need of leveraging large generative language model for multimodal applications. Recently, a few notable multimodal models (e.g., BLIP, Flamingo, KOSMOS, PaLM-E, LLaVA, Visual ChatGPT, GPT-4, etc.) with a combination of large language models significantly enhanced their understanding and generate more accurate and nuanced responses. The workshop will provide an opportunity for researchers, practitioners, and industry professionals to explore the latest trends and best practices in the field of multimodal applications of large generative models. The workshop will also focus on exploring the challenges and opportunities of integrating large language models with other AI technologies such as computer vision and speech recognition.

Xu Tan 0003

With the development of deep learning, multimedia data generation (e.g., image generation, audio synthesis, music composition, and video generation) has attracted a lot of attention. Deep learning methods for data generation usually build a mapping from source condition X to target data Y. The target Y (e.g., image, speech, music, video) is usually high-dimensional and complex, and contains rich information not exist in source data, which hinders the effective and efficient learning on the source-target mapping. Representation learning has achieved rapid progress in the past decade, which is beneficial for data understanding tasks. However, traditional representation learning cannot address the challenges faced by multimedia data generation tasks. This tutorial revisits the learning paradigms for data generation and introduces a paradigm called regeneration learning that can improve the effectiveness and efficiency of multimedia data generation. We show that a variety of tasks in multimedia data generation (e.g., image generation, speech synthesis, music composition, video generation) can benefit from this regeneration learning paradigm, and a lot of recent popular data generation models (e.g., DALL-E 1/2, Stable Diffusion, AudioLM, NaturalSpeech 2, MusicLM) can be covered by this learning paradigm.

Aaron Keesing, Yun Sing Koh, Vithya Yogarajan, Michael Witbrock

Many software packages and toolkits have been developed for machine learning, in particular for natural language processing and automatic speech recognition. However, there are few software packages designed for emotion recognition. Emotion datasets have diverse structures and annotations, and feature extractors often have different interfaces, which requires writing code specific to each interface. To improve the standardisation and reproducibility of emotion recognition research, we present the Emotion Recognition ToolKit (ERTK), a Python library for emotion recognition. ERTK comprises processing scripts for emotion datasets, standard interfaces to feature extractors, and a framework for defining experiments with declarative configuration files. ERTK is modular and extensible, which allows for easily incorporating additional models and processors. The current version of ERTK focuses on emotional speech, however, the library is modular and can be easily extended to other modalities, which we plan for future releases. ERTK is open-source and available from GitHub: https://github.com/Strong-AI-Lab/emotion.

Björn W. Schuller, Anton Batliner, Shahin Amiriparian, Alexander Barnhill, Maurice Gerczuk, Andreas Triantafyllopoulos, Alice E. Baird, Panagiotis Tzirakis, Chris Gagne 0001, Alan S. Cowen 等

The ACM Multimedia 2023 Computational Paralinguistics Challenge addresses two different problems for the first time in a research competition under well-defined conditions: In the Emotion Share Sub-Challenge, a regression on speech has to be made; and in the Requests Sub-Challenges, requests and complaints need to be detected. We describe the Sub-Challenges, baseline feature extraction, and classifiers based on the 'usual' ComPaRE features, the auDeep toolkit, and deep feature extraction from pre-trained CNNs using the DeepSpectRum toolkit; in addition, wav2vec2 models are used.

Daoming Zong, Chaoyue Ding, Baoxiang Li, Dinghao Zhou, Jiakui Li, Ken Zheng, Qunyan Zhou 0002

In this paper, we present the solutions to the MER-MULTI and MER-NOISE sub-challenges of the Multimodal Emotion Recognition Challenge (MER 2023). For the tasks MER-MULTI and MER-NOISE, participants are required to recognize both discrete and dimensional emotions. Particularly, in MER-NOISE, the test videos are corrupted with noise, necessitating the consideration of modality robustness. Our empirical findings indicate that different modalities contribute differently to the tasks, with a significant impact from the audio and visual modalities, while the text modality plays a weaker role in emotion prediction. To facilitate subsequent multimodal fusion, and considering that language information is implicitly embedded in large pre-trained speech models, we have made the deliberate choice to abandon the text modality and solely utilize visual and acoustic modalities for these sub-challenges. To address the potential underfitting of individual modalities during multimodal training, we propose to jointly train all modalities via a weighted blending of supervision signals. Furthermore, to enhance the robustness of our model, we employ a range of data augmentation techniques at the image level, waveform level, and spectrogram level. Experimental results show that our model ranks 1st in both MER-MULTI (0.7005) and MER-NOISE (0.6846) sub-challenges, validating the effectiveness of our method. Our code is publicly available at https://github.com/dingchaoyue/Multimodal-Emotion-Recognition-MER-and-MuSe-2023-Challenges.

Payal Mohapatra, Akash Pandey, Yueyuan Sui, Qi Zhu 0002

Human emotion understanding is pivotal in making conversational technology mainstream. We view speech emotion understanding as a perception task which is a more realistic setting. With varying contexts (languages, demographics etc.) different share of people perceive the same speech segment as a non-unanimous emotion. As part of the ACM Multimedia 2023 Computational Paralinguistics ChallengE (ComParE) in the EMotion Share track, we leverage their rich dataset of multilingual speakers and multi-label regression target of 'emotion share' or perception of that emotion. We demonstrate that the training scheme of different foundation models dictates their effectiveness for tasks beyond speech recognition, especially for non-semantic speech tasks like emotion understanding. This is a very complex task due to multilingual speakers, variability in the target labels, and inherent imbalance in the regression dataset. Our results show that HuBERT-Large with a self-attention-based light-weight sequence model provides 4.6% improvement over the reported baseline.

Nicolae-Catalin Ristea, Radu Tudor Ionescu

We propose a novel cascaded cross-modal transformer (CCMT) that combines speech and text transcripts to detect customer requests and complaints in phone conversations. Our approach leverages a multimodal paradigm by transcribing the speech using automatic speech recognition (ASR) models and translating the transcripts into different languages. Subsequently, we combine language-specific BERT-based models with Wav2Vec2.0 audio features in a novel cascaded cross-attention transformer model. We apply our system to the Requests Sub-Challenge of the ACM Multimedia 2023 Computational Paralinguistics Challenge, reaching unweighted average recalls (UAR) of 65.41% and 85.87% for the complaint and request classes, respectively.

Yuxuan Zhang, Lei Liu, Li Liu

Cued Speech (CS) is a visual coding tool to encode spoken languages at the phonetic level, which combines lip-reading and hand gestures to effectively assist communication among people with hearing impairments. The Automatic CS Recognition (ACSR) task aims to recognize CS videos into linguistic texts, which involves both lips and hands as two distinct modalities conveying complementary information. However, the traditional centralized training approach poses potential privacy risks due to the use of facial and gesture videos in CS data. To address this issue, we propose a new Federated Cued Speech Recognition (FedCSR) framework to train an ACSR model over the decentralized CS data without sharing private information. In particular, a mutual knowledge distillation method is proposed to maintain cross-modal semantic consistency of the Non-IID CS data, which ensures learning a unified feature space for linguistic and visual information. On the server side, a globally shared linguistic model is trained to capture the long-term dependencies in the text sentences, which is aligned with the visual information from the local clients via visual-to-linguistic distillation. On the client side, the visual model of each client is trained with its own local data, assisted by linguistic-to-visual distillation treating the linguistic model as the teacher. To the best of our knowledge, this is the first approach to consider the federated ACSR task for privacy protection. Experimental results on the Chinese CS dataset with multiple cuers demonstrate that our approach outperforms both mainstream federated learning baselines and existing centralized state-of-the-art ACSR methods, achieving 9.7% performance improvement for character error rate (CER) and 15.0% for word error rate (WER). The Chinese CS dataset and our code will be open-sourced.

Yanzhen Ren, Hongcheng Zhu, Liming Zhai, Zongkun Sun, Rubing Shen, Lina Wang 0001

Voice conversion (VC), as a voice style transfer technology, is becoming increasingly prevalent while raising serious concerns about its illegal use. Proactively tracing the origins of VC-generated speeches, i.e., speaker traceability, can prevent the misuse of VC, but unfortunately has not been extensively studied. In this paper, we are the first to investigate the speaker traceability for VC and propose a traceable VC framework named VoxTracer. Our VoxTracer is similar to but beyond the paradigm of audio watermarking. We first use unique speaker embedding to represent speaker identity. Then we design a VAE-Glow structure, in which the hiding process imperceptibly integrates the source speaker identity into the VC, and the tracing process accurately recovers the source speaker identity and even the source speech in spite of severe speech quality degradation. To address the speech mismatch between the hiding and tracing processes affected by different distortions, we also adopt an asynchronous training strategy to optimize the VAE-Glow models. The VoxTracer is versatile enough to be applied to arbitrary VC methods and popular audio coding standards. Extensive experiments demonstrate that the VoxTracer achieves not only high imperceptibility in hiding, but also nearly 100% tracing accuracy against various types of audio lossy compressions (AAC, MP3, Opus and SILK) with a broad range of bitrates (16 kbps - 128 kbps) even in a very short time duration (0.74s). Our source code is available at https://github.com/hongchengzhu/VoxTracer.

Xuan Hai, Xin Liu 0050, Yuan Tan 0003, Qingguo Zhou

Voice is a vital medium for transmitting information. The advancement of speech synthesis technology has resulted in high-quality synthesized voices indistinguishable from human ears. These fake voices have been widely used in natural Deepfake production and other malicious activities, raising serious concerns regarding security and privacy. To deal with this situation, there have been many studies working on detecting fake voices and reporting excellent performance. However, is the story really over? In this paper, we propose SiFDetectCracker, a black-box adversarial attack framework based on Speaker-Irrelative Features (SiFs) against fake voice detection. We select background noise and mute parts before and after the speaker's voice as the primary attack features. By modifying these features in synthesized speech, the fake speech detector will make a misjudgment. Experiments show that SiFDetectCracker achieved a success rate of more than 80% in bypassing existing state-of-the-art fake voice detection systems. We also conducted several experiments to evaluate our attack approach's transferability and activation factor.

Zhicong Zheng, Xinfeng Li, Chen Yan 0001, Xiaoyu Ji 0001, Wenyuan Xu 0001

Backdoor Attacks have been shown to pose significant threats to automatic speech recognition systems (ASRs). Existing success largely assumes backdoor triggering in the digital domain, or the victim will not notice the presence of triggering sounds in the physical domain. However, in practical victim-present scenarios, the over-the-air distortion of the backdoor trigger and the victim awareness raised by its audibility may invalidate such attacks. In this paper, we propose SMA, an inaudible grey-box backdoor attack that can be generalized to real-world scenarios where victims are present by exploiting both the vulnerability of microphones and neural networks. Specifically, we utilize the nonlinear effects of microphones to inject an inaudible ultrasonic trigger. To accurately characterize the microphone response to the crafted ultrasound, we construct a novel nonlinear transfer function for effective optimization. We also design optimization objectives to ensure triggers' robustness in the physical world and transferability on unseen ASR models. In practice, SMA can bypass the microphone's built-in filters and human perception, activating the implanted trigger in the ASRs inaudibly, regardless of whether the user is speaking. Extensive experiments show that the attack success rate of SMA can reach nearly 100% in the digital domain and over 85% against most microphones in the physical domains by only poisoning about 0.5% of the training audio dataset. Moreover, our attack can resist typical defense countermeasures to backdoor attacks.

Songlin Yang, Wei Wang 0025, Jun Ling, Bo Peng 0002, Xu Tan 0003, Jing Dong 0003

Talking-head video editing aims to efficiently insert, delete, and substitute the word of a pre-recorded video through a text transcript editor. The key challenge for this task is obtaining an editing model that generates new talking-head video clips which simultaneously have accurate lip synchronization and motion smoothness. Previous approaches, including 3DMM-based (3D Morphable Model) methods and NeRF-based (Neural Radiance Field) methods, are sub-optimal in that they either require minutes of source videos and days of training time or lack the disentangled control of verbal (e.g., lip motion) and non-verbal (e.g., head pose and expression) representations for video clip insertion. In this work, we fully utilize the video context to design a novel framework for talking-head video editing, which achieves efficiency, disentangled motion control, and sequential smoothness. Specifically, we decompose this framework to motion prediction and motion-conditioned rendering: (1) We first design an animation prediction module that efficiently obtains smooth and lip-sync motion sequences conditioned on the driven speech. This module adopts a non-autoregressive network to obtain context prior and improve the prediction efficiency, and it learns a speech-animation mapping prior with better generalization to novel speech from a multi-identity video dataset. (2) We then introduce a neural rendering module to synthesize the photo-realistic and full-head video frames given the predicted motion sequence. This module adopts a pre-trained head topology and uses only few frames for efficient fine-tuning to obtain a person-specific rendering model. Extensive experiments demonstrate that our method efficiently achieves smoother editing results with higher image quality and lip accuracy using less data than previous methods.

Haozhe Wu, Songtao Zhou, Jia Jia 0001, Junliang Xing, Qi Wen, Xiang Wen

Speech-driven 3D face animation poses significant challenges due to the intricacy and variability inherent in human facial movements. This paper emphasizes the importance of considering both the composite and regional natures of facial movements in speech-driven 3D face animation. The composite nature pertains to how speech-independent factors globally modulate speech-driven facial movements along the temporal dimension. Meanwhile, the regional nature alludes to the notion that facial movements are not globally correlated but are actuated by local musculature along the spatial dimension. It is thus indispensable to incorporate both natures for engendering vivid animation. To address the composite nature, we introduce an adaptive modulation module that employs arbitrary facial movements to dynamically adjust speech-driven facial movements across frames on a global scale. To accommodate the regional nature, our approach ensures that each constituent of the facial features for every frame focuses on the local spatial movements of 3D faces. Moreover, we present a non-autoregressive backbone for translating audio to 3D facial movements, which maintains high-frequency nuances of facial movements and facilitates efficient inference. Comprehensive experiments and user studies demonstrate that our method surpasses contemporary state-of-the-art approaches both qualitatively and quantitatively.