2003-예방원년, 소아과 의사의 역할

여성사 쓰기의 가능성과 현주소, 한국여성연구소 여성사연구실 지음, 청년사 1999

In this work, we present a hybrid CTC/Attention model based on a ResNet-18 and Convolution-augmented transformer (Conformer), that can be trained in an end-to-end manner. In particular, the audio and visual encoders learn to extract features directly from raw pixels and audio waveforms, respectively, which are then fed to conformers and then fusion takes place via a Multi-Layer Perceptron (MLP). The model learns to recognise characters using a combination of CTC and an attention mechanism. We show that end-to-end training, instead of using pre-computed visual features which is common in the literature, the use of a conformer, instead of a recurrent network, and the use of a transformer-based language model, significantly improve the performance of our model. We present results on the largest publicly available datasets for sentence-level speech recognition, Lip Reading Sentences 2 (LRS2) and Lip Reading Sentences 3 (LRS3), respectively. The results show that our proposed models raise the state-of-the-art performance by a large margin in audio-only, visual-only, and audio-visual experiments.

End-To-End Audio-Visual Speech Recognition with Conformers

Recent works in speech recognition rely either on connectionist temporal classification (CTC) or sequence-to-sequence models for character-level recognition. CTC assumes conditional independence of individual characters, whereas attention-based models can provide nonsequential alignments. Therefore, we could use a CTC loss in combination with an attention-based model in order to force monotonic alignments and at the same time get rid of the conditional independence assumption. In this paper, we use the recently proposed hybrid CTC/attention architecture for audio-visual recognition of speech in-the-wild. To the best of our knowledge, this is the first time that such a hybrid architecture architecture is used for audio-visual recognition of speech. We use the LRS2 database and show that the proposed audio-visual model leads to an 1.3% absolute decrease in word error rate over the audio-only model and achieves the new state-of-the-art performance on LRS2 database (7% word error rate). We also observe that the audio-visual model significantly outperforms the audio-based model (up to 32.9% absolute improvement in word error rate) for several different types of noise as the signal-to-noise ratio decreases.

Audio-Visual Speech Recognition with a Hybrid CTC/Attention Architecture

Survey on automatic lip-reading in the era of deep learning

Automatic speech recognition can potentially benefit from the lip motion patterns, complementing acoustic speech to improve the overall recognition performance, particularly in noise. In this paper we propose an audio-visual fusion strategy that goes beyond simple feature concatenation and learns to automatically align the two modalities, leading to enhanced representations which increase the recognition accuracy in both clean and noisy conditions. We test our strategy on the TCD-TIMIT and LRS2 datasets, designed for large vocabulary continuous speech recognition, applying three types of noise at different power ratios. We also exploit state of the art Sequence-to-Sequence architectures, showing that our method can be easily integrated. Results show relative improvements from 7% up to 30% on TCD-TIMIT over the acoustic modality alone, depending on the acoustic noise level. We anticipate that the fusion strategy can easily generalise to many other multimodal tasks which involve correlated modalities.

/pdf/attention-based-audio-visual-fusion-for-robust-automatic-3b5uw9nzx5.pdf

Attention-based Audio-Visual Fusion for Robust Automatic Speech Recognition

Audio-Visual Speech Recognition (AVSR) seeks to model, and thereby exploit, the dynamic relationship between a human voice and the corresponding mouth movements. A recently proposed multimodal fusion strategy, AV Align, based on state-of-the-art sequence to sequence neural networks, attempts to model this relationship by explicitly aligning the acoustic and visual representations of speech. This study investigates the inner workings of AV Align and visualises the audio-visual alignment patterns. Our experiments are performed on two of the largest publicly available AVSR datasets, TCD-TIMIT and LRS2. We find that AV Align learns to align acoustic and visual representations of speech at the frame level on TCD-TIMIT in a generally monotonic pattern. We also determine the cause of initially seeing no improvement over audio-only speech recognition on the more challenging LRS2. We propose a regularisation method which involves predicting lip-related Action Units from visual representations. Our regularisation method leads to better exploitation of the visual modality, with performance improvements between 7% and 30% depending on the noise level. Furthermore, we show that the alternative Watch, Listen, Attend, and Spell network is affected by the same problem as AV Align, and that our proposed approach can effectively help it learn visual representations. Our findings validate the suitability of the regularisation method to AVSR and encourage researchers to rethink the multimodal convergence problem when having one dominant modality.

How to Teach DNNs to Pay Attention to the Visual Modality in Speech Recognition

Automatic speech recognition can potentially benefit from the lip motion patterns, complementing acoustic speech to improve the overall recognition performance, particularly in noise In this paper we propose an audio-visual fusion strategy that goes beyond simple feature concatenation and learns to automatically align the two modalities, leading to enhanced representations which increase the recognition accuracy in both clean and noisy conditions We test our strategy on the TCD-TIMIT and LRS2 datasets, designed for large vocabulary continuous speech recognition, applying three types of noise at different power ratios We also exploit state of the art Sequence-to-Sequence architectures, showing that our method can be easily integrated Results show relative improvements from 7% up to 30% on TCD-TIMIT over the acoustic modality alone, depending on the acoustic noise level We anticipate that the fusion strategy can easily generalise to many other multimodal tasks which involve correlated modalities Code available online on GitHub: this https URL

/pdf/attention-based-audio-visual-fusion-for-robust-automatic-2jd6ddomd2.pdf

Automatic lipreading has major potential impact for speech recognition, supplementing and complementing the acoustic modality. Most attempts at lipreading have been performed on small vocabulary tasks, due to a shortfall of appropriate audio-visual datasets. In this work we use the publicly available TCD-TIMIT database, designed for large vocabulary continuous audio-visual speech recognition. We compare the viseme recognition performance of the most widely used features for lipreading, Discrete Cosine Transform (DCT) and Active Appearance Models (AAM), in a traditional Hidden Markov Model (HMM) framework. We also exploit recent advances in AAM fitting. We found the DCT to outperform AAM by more than 6% for a viseme recognition task with 56 speakers. The overall accuracy of the DCT is quite low (32-34%). We conclude that a fundamental rethink of the modelling of visual features may be needed for this task.

Towards Lipreading Sentences with Active Appearance Models

Audio-Visual Speech Recognition (AVSR) seeks to model, and thereby exploit, the dynamic relationship between a human voice and the corresponding mouth movements. A recently proposed multimodal fusion strategy, AV Align , based on state-of-the-art sequence to sequence neural networks, attempts to model this relationship by explicitly aligning the acoustic and visual representations of speech. This study investigates the inner workings of AV Align and visualises the audio-visual alignment patterns. Our experiments are performed on two of the largest publicly available AVSR datasets, TCD-TIMIT and LRS2. We find that AV Align learns to align acoustic and visual representations of speech at the frame level on TCD-TIMIT in a generally monotonic pattern. We also determine the cause of initially seeing no improvement over audio-only speech recognition on the more challenging LRS2. We propose a regularisation method which involves predicting lip-related Action Units from visual representations. Our regularisation method leads to better exploitation of the visual modality, with performance improvements between 7% and 30% depending on the noise level. Furthermore, we show that the alternative Watch, Listen, Attend, and Spell network is affected by the same problem as AV Align , and that our proposed approach can effectively help it learn visual representations. Our findings validate the suitability of the regularisation method to AVSR and encourage researchers to rethink the multimodal convergence problem when having one dominant modality.

George Sterpu

Papers

Attention-based Audio-Visual Fusion for Robust Automatic Speech Recognition

How to Teach DNNs to Pay Attention to the Visual Modality in Speech Recognition

Attention-based Audio-Visual Fusion for Robust Automatic Speech Recognition

Towards Lipreading Sentences with Active Appearance Models

How to Teach DNNs to Pay Attention to the Visual Modality in Speech Recognition