Visual speech information from the speaker's mouth region has been successfully shown to improve noise robustness of automatic speech recognizers, thus promising to extend their usability in the human computer interface. In this paper, we review the main components of audiovisual automatic speech recognition (ASR) and present novel contributions in two main areas: first, the visual front-end design, based on a cascade of linear image transforms of an appropriate video region of interest, and subsequently, audiovisual speech integration. On the latter topic, we discuss new work on feature and decision fusion combination, the modeling of audiovisual speech asynchrony, and incorporating modality reliability estimates to the bimodal recognition process. We also briefly touch upon the issue of audiovisual adaptation. We apply our algorithms to three multisubject bimodal databases, ranging from small- to large-vocabulary recognition tasks, recorded in both visually controlled and challenging environments. Our experiments demonstrate that the visual modality improves ASR over all conditions and data considered, though less so for visually challenging environments and large vocabulary tasks.

/pdf/recent-advances-in-the-automatic-recognition-of-audiovisual-30c21z0cpc.pdf

Recent advances in the automatic recognition of audiovisual speech

We have made significant progress in automatic speech recognition (ASR) for well-defined applications like dictation and medium vocabulary transaction processing tasks in relatively controlled environments. However, ASR performance has yet to reach the level required for speech to become a truly pervasive user interface. Indeed, even in “clean” acoustic environments, and for a variety of tasks, state of the art ASR system performance lags human speech perception by up to an order of magnitude (Lippmann, 1997). In addition, current systems are quite sensitive to channel, environment, and style of speech variations. A number of techniques for improving ASR robustness have met limited success in severely degraded environments, mismatched to system training (Ghitza, 1986; Nadas et al., 1989; Juang, 1991; Liu et al., 1993; Hermansky and Morgan, 1994; Neti, 1994; Gales, 1997; Jiang et al., 2001). Clearly, novel, non-traditional approaches, that use orthogonal sources of information to the acoustic input, are needed to achieve ASR performance closer to the human speech perception level, and robust enough to be deployable in field applications. Visual speech is the most promising source of additional speech information, and it is obviously not affected by the acoustic environment and noise.

Audio-Visual Automatic Speech Recognition: An Overview

Microphones and cameras have been extensively used to observe and detect human activity and to facilitate natural modes of interaction between humans and intelligent systems. Human brain processes the audio and video modalities, extracting complementary and robust information from them. Intelligent systems with audiovisual sensors should be capable of achieving similar goals. The audiovisual information fusion strategy is a key component in designing such systems. In this paper, we exclusively survey the fusion techniques used in various audiovisual information fusion tasks. The fusion strategy used tends to depend mainly on the model, probabilistic or otherwise, used in the particular task to process sensory information to obtain higher level semantic information. The models themselves are task oriented. In this paper, we describe the fusion strategies and the corresponding models used in audiovisual tasks such as speech recognition, tracking, biometrics, affective state recognition, and meeting scene analysis. We also review the challenges and existing solutions and also unresolved or partially resolved issues in these fields. Specifically, we discuss established and upcoming work in hierarchical fusion strategies and cross-modal learning techniques, identifying these as critical areas of research in the future development of intelligent systems.

http://cvrr.ucsd.edu/publications/2010/ShivappaTrivediRao_PIEEE2010.pdf

Audiovisual Information Fusion in Human–Computer Interfaces and Intelligent Environments: A Survey

Historically, the development of computer interfaces has been a technology-driven phenomenon. However, new multimodal interfaces are composed of recognition-based technologies that must interpret human speech, gesture, gaze, movement patterns, and other complex natural behaviors, which involve highly automatized skills that are not under full conscious control. As a result, it now is widely acknowledged that multimodal interface design requires modeling of the modality-centered behavior and integration patterns upon which multimodal systems aim to build. This paper summarizes research on the cognitive science foundations of multimodal interaction, and on the essential role that user-centered modeling has played in prototyping, guiding, and evaluating the design of next-generation multimodal interfaces. In particular, it discusses the properties of different modalities and the information content they carry, the unique features of multimodal language and its processability, as well as when users are likely to interact multimodally and how their multimodal input is integrated and synchronized. It also reviews research on typical performance and linguistic efficiencies associated with multimodal interaction, and on the user-centered reasons why multimodal interaction minimizes errors and expedites error handling. In addition, this paper describes the important role that selective methodologies and evaluation metrics have played in shaping next-generation multimodal systems, and it concludes by highlighting future directions for designing a new class of adaptive multimodal-multisensor interfaces.

User-centered modeling and evaluation of multimodal interfaces

The integration of audio and visual information improves speech recognition performance, specially in the presence of noise. In these circumstances it is necessary to introduce audio and visual weights to control the contribution of each modality to the recognition task. We present a method to set the value of the weights associated to each stream according to their reliability for speech recognition, allowing them to change with time and adapt to different noise and working conditions. Our dynamic weights are derived from several measures of the stream reliability, some specific to speech processing and others inherent to any classification task, and take into account the special role of silence detection in the definition of audio and visual weights. In this paper, we propose a new confidence measure, compare it to existing ones, and point out the importance of the correct detection of silence utterances in the definition of the weighting system. Experimental results support our main contribution: the inclusion of a voice activity detector in the weighting scheme improves speech recognition over different system architectures and confidence measures, leading to an increase in performance more relevant than any difference between the proposed confidence measures.

On Dynamic Stream Weighting for Audio-Visual Speech Recognition

ICSLP2000: the 6th International Conference on Spoken Language Processing, October 16-20, 2000, Beijing, China.

Hidetoshi Ito

Papers

Stream Weight Optimization of Speech and Lip Image Sequence for Audio-Visual Speech Recognition