Microphone arrays have been used in various applications to capture conversations, such as in meetings and teleconferences. In many cases, the microphone and likely source locations are known a priori, and calculating beamforming filters is therefore straightforward. In ad-hoc situations, however, when the microphones have not been systematically positioned, this information is not available and beamforming must be achieved blindly. In achieving this, a commonly neglected issue is whether it is optimal to use all of the available microphones, or only an advantageous subset of these. This paper commences by reviewing different approaches to blind beamforming, characterizing them by the way they estimate the signal propagation vector and the spatial coherence of noise in the absence of prior knowledge of microphone and speaker locations. Following this, a novel clustered approach to blind beamforming is motivated and developed. Without using any prior geometrical information, microphones are first grouped into localized clusters, which are then ranked according to their relative distance from a speaker. Beamforming is then performed using either the closest microphone cluster, or a weighted combination of clusters. The clustered algorithms are compared to the full set of microphones in experiments on a database recorded on different ad-hoc array geometries. These experiments evaluate the methods in terms of signal enhancement as well as performance on a large vocabulary speech recognition task.

For several reasons, the Fourier phase domain is less favored than the magnitude domain in signal processing and modeling of speech. To correctly analyze the phase, several factors must be considered and compensated, including the effect of the step size, windowing function and other processing parameters. Building on a review of these factors, this paper investigates a spectral representation based on the Instantaneous Frequency Deviation, but in which the step size between processing frames is used in calculating phase changes, rather than the traditional single sample interval. Reflecting these longer intervals, the term delta-phase spectrum is used to distinguish this from instantaneous derivatives. Experiments show that mel-frequency cepstral coefficients features derived from the delta-phase spectrum (termed Mel-Frequency delta-phase features) can produce broadly similar performance to equivalent magnitude domain features for both voice activity detection and speaker recognition tasks. Further, it is shown that the fusion of the magnitude and phase representations yields performance benefits over either in isolation.

This paper presents a method of voice activity detection (VAD) suitable for high noise scenarios, based on the fusion of two complementary systems. The first system uses a proposed non-Gaussianity score (NGS) feature based on normal probability testing. The second system employs a histogram distance score (HDS) feature that detects changes in the signal through conducting a template-based similarity measure between adjacent frames. The decision outputs by the two systems are then merged using an open-by-reconstruction fusion stage. Accuracy of the proposed method was compared to several baseline VAD methods on a database created using real recordings of a variety of high-noise environments.

This paper proposes a clustered approach for blind beamfoming from ad-hoc microphone arrays. In such arrangements, microphone placement is arbitrary and the speaker may be close to one, all or a subset of microphones at a given time. Practical issues with such a configuration mean that some microphones might be better discarded due to poor input signal to noise ratio (SNR) or undesirable spatial aliasing effects from large inter-element spacings when beamforming. Large inter-microphone spacings may also lead to inaccuracies in delay estimation during blind beamforming. In such situations, using a cluster of microphones (ie, a sub-array), closely located both to each other and to the desired speech source, may provide more robust enhancement than the full array. This paper proposes a method for blind clustering of microphones based on the magnitude square coherence function, and evaluates the method on a database recorded using various ad-hoc microphone arrangements.

This paper investigates robustness to uncertain microphone placements in an array beamformer front-end to a speech recognition system. There are two general approaches to handling the placement uncertainty: using the approximately known geometry in a robust beamforming technique, or using techniques that require no prior knowledge of geometry. Experiments in this paper compare the robustness of different techniques for both of these approaches in terms of speech recognition accuracy. To benefit from existing microphone array speech recognition data corpora for experimentation, microphone placement uncertainty is simulated by introducing random perturbations in the assumed geometry. Experimental results show that robust beamforming yields stable performance to a certain degree of placement error, but thereafter techniques such as automatic calibration are beneficial.

This correspondence presents a microphone array shape calibration procedure for diffuse noise environments. The procedure estimates intermicrophone distances by fitting the measured noise coherence with its theoretical model and then estimates the array geometry using classical multidimensional scaling. The technique is validated on noise recordings from two office environments.

Cancer staging provides a basis for planning clinical management, but also allows for meaningful analysis of cancer outcomes and evaluation of cancer care services. Despite this, stage data in cancer registries is often incomplete, inaccurate, or simply not collected. This article describes a prototype software system (Cancer Stage Interpretation System, CSIS) that automatically extracts cancer staging information from medical reports. The system uses text classification techniques to train support vector machines (SVMs) to extract elements of stage listed in cancer staging guidelines. When processing new reports, CSIS identifies sentences relevant to the staging decision, and subsequently assigns the most likely stage. The system was developed using a database of staging data and pathology reports for 710 lung cancer patients, then validated in an independent set of 179 patients against pathologic stage assigned by two independent pathologists. CSIS achieved overall accuracy of 74% for tumor (T) staging and 87% for node (N) staging, and errors were observed to mirror disagreements between human experts.

Tracking speakers in multiparty conversations constitutes a fundamental task for automatic meeting analysis. In this paper, we present a novel probabilistic approach to jointly track the location and speaking activity of multiple speakers in a multisensor meeting room, equipped with a small microphone array and multiple uncalibrated cameras. Our framework is based on a mixed-state dynamic graphical model defined on a multiperson state-space, which includes the explicit definition of a proximity-based interaction model. The model integrates audiovisual (AV) data through a novel observation model. Audio observations are derived from a source localization algorithm. Visual observations are based on models of the shape and spatial structure of human heads. Approximate inference in our model, needed given its complexity, is performed with a Markov Chain Monte Carlo particle filter (MCMC-PF), which results in high sampling efficiency. We present results-based on an objective evaluation procedure-that show that our framework 1) is capable of locating and tracking the position and speaking activity of multiple meeting participants engaged in real conversations with good accuracy, 2) can deal with cases of visual clutter and occlusion, and 3) significantly outperforms a traditional sampling-based approach

This paper addresses the problem of distant speech acquisition in multiparty meetings, using multiple microphones and cameras. Microphone array beamforming techniques present a potential alternative to close-talking microphones by providing speech enhancement through spatial filtering. Beamforming techniques, however, rely on knowledge of the speaker location. In this paper, we present an integrated approach, in which an audio– visual multiperson tracker is used to track active speakers with high accuracy. Speech enhancement is then achieved using microphone array beamforming followed by a novel postfiltering stage. Finally, speech recognition is performed to evaluate the quality of the enhanced speech signal. The approach is evaluated on data recorded in a real meeting room for stationary speaker, moving speaker, and overlapping speech scenarios. The results show that the speech enhancement and recognition performance achieved using our approach are significantly better than a single table-top microphone and are comparable to a lapel microphone for some of the scenarios. The results also indicate that the audio–visual-based system performs significantly better than audio-only system, both in terms of enhancement and recognition. This reveals that the accurate speaker tracking provided by the audio–visual sensor array proved beneficial to improve the recognition performance in a microphone array-based speech recognition system.

We address the problem of recognizing sequences of human interaction patterns in meetings, with the goal of structuring them in semantic terms. The investigated patterns are inherently group-based (defined by the individual activities of meeting participants, and their interplay), and multimodal (as captured by cameras and microphones). By defining a proper set of individual actions, group actions can be modeled as a two-layer process, one that models basic individual activities from low-level audio-visual (AV) features,and another one that models the interactions. We propose a two-layer hidden Markov model (HMM) framework that implements such concept in a principled manner, and that has advantages over previous works. First, by decomposing the problem hierarchically, learning is performed on low-dimensional observation spaces, which results in simpler models. Second, our framework is easier to interpret, as both individual and group actions have a clear meaning, and thus easier to improve. Third, different HMMs can be used in each layer, to better reflect the nature of each subproblem. Our framework is general and extensible, and we illustrate it with a set of eight group actions, using a public 5-hour meeting corpus. Experiments and comparison with a single-layer HMM baseline system show its validity.

We address the problem of temporal unusual event detection. Unusual events are characterized by a number of features (rarity, unexpectedness, and relevance) that limit the application of traditional supervised model-based approaches. We propose a semi-supervised adapted hidden Markov model (HMM) framework, in which usual event models are first learned from a large amount of (commonly available) training data, while unusual event models are learned by Bayesian adaptation in an unsupervised manner. The proposed framework has an iterative structure, which adapts a new unusual event model at each iteration. We show that such a framework can address problems due to the scarcity of training data and the difficulty in pre-defining unusual events. Experiments on audio, visual, and audiovisual data streams illustrate its effectiveness, compared with both supervised and unsupervised baseline methods.

Finding relevant segments in meeting recordings is important for summarization, browsing, and retrieval purposes. In this paper, we define relevance as the interest-level that meeting participants manifest as a group during the course of their interaction (as perceived by an external observer), and investigate the automatic detection of segments of high-interest from audio-visual cues. This is motivated by the assumption that there is a relationship between segments of interest to participants, and those of interest to the end user, e.g. of a meeting browser. We first address the problem of human annotation of group interest-level. On a 50-meeting corpus, recorded in a room equipped with multiple cameras and microphones, we found that the annotations generated by multiple people exhibit a good degree of consistency, providing a stable ground-truth for automatic methods. For the automatic detection of high-interest segments, we investigate a methodology based on hidden Markov models (HMM) and a number of audio and visual features. Single- and multi-stream approaches were studied. Using precision and recall as performance measures, the results suggest that the automatic detection of group interest-level is promising, and that while audio in general constitutes the predominant modality in meetings, the use of a multi-modal approach is beneficial.

Close-talk headset microphones have been traditionally used for speech acquisition in a number of applications, as they naturally provide a higher signal-to-noise ratio - needed for recognition tasks than single distant microphones. However, in multi-party conversational settings like meetings, microphone arrays represent an important alternative to close-talking microphones, as they allow for localisation and tracking of speakers and signal-independent enhancement, while providing a non-intrusive, hands-free operation mode. In this article, we investigate the use of an audio-visual sensor array, composed of a small table-top microphone array and a set of cameras, for speaker tracking and speech enhancement in meetings. Our methodology first fuses audio and video for person tracking, and then integrates the output of the tracker with a beamformer for speech enhancement. We compare and discuss the features of the resulting speech signal with respect to that obtained from single close-talking and table-top microphones.

Tracking speakers in multiparty conversations constitutes a fundamental task for automatic meeting analysis. In this paper, we present a probabilistic approach to jointly track the location and speaking activity of multiple speakers in a multisensor meeting room, equipped with a small microphone array and multiple uncalibrated cameras. Our framework is based on a mixed-state dynamic graphical model defined on a multiperson state-space, which includes the explicit definition of a proximity-based interaction model. The model integrates audio-visual (AV) data through a novel observation model. Audio observations are derived from a source localization algorithm. Visual observations are based on models of the shape and spatial structure of human heads. Approximate inference in our model, needed given its complexity, is performed with a Markov Chain Monte Carlo particle filter (MCMC-PF), which results in high sampling efficiency. We present results -based on an objective evaluation procedure-that show that our framework (1) is capable of locating and tracking the position and speaking activity of multiple meeting participants engaged in real conversations with good accuracy; (2) can deal with cases of visual clutter and partial occlusion; and (3) significantly outperforms a traditional sampling-based approach.

This paper investigates the recognition of group actions in meetings. A framework is employed in which group actions result from the interactions of the individual participants. The group actions are modeled using different HMM-based approaches, where the observations are provided by a set of audiovisual features monitoring the actions of individuals. Experiments demonstrate the importance of taking interactions into account in modeling the group actions. It is also shown that the visual modality contains useful information, even for predominantly audio-based events, motivating a multimodal approach to meeting analysis.

The paper presents a new approach toward automatic annotation of meetings in terms of speaker identities and their locations. This is achieved by segmenting the audio recordings using two independent sources of information: magnitude spectrum analysis and sound source localization. We combine the two in an appropriate HMM framework. There are three main advantages of this approach. First, it is completely unsupervised, i.e. speaker identities and number of speakers and locations are automatically inferred. Second, it is threshold-free, i.e. the decisions are made without the need of a threshold value which generally requires an additional development dataset. The third advantage is that the joint segmentation improves over the speaker segmentation derived using only acoustic features. Experiments on a series of meetings recorded in the IDIAP smart meeting room demonstrate the effectiveness of this approach.

This paper presents overview of an online audio indexing system, which creates a searchable index of speech content embedded in digitized audio files. This system is based on our recently proposed offline audio segmentation techniques. As the data arrives continuously, the system first finds boundaries of the acoustically homogenous segments. Next, each of these segments is classified as speech, music or {\\it mixture} classes, where mixtures are defined as regions where speech and other non-speech sounds are present simultaneously and noticeably. The speech segments are then clustered together to provide consistent speaker labels. The speech and mixture segments are converted to text via an ASR system. The resulting words are time-stamped together with other metadata information (speaker identity, speech confidence score) in an XML file to rapidly identify and access target segments. In this paper, we analyze the performance at each stage of this audio indexing system and also compare it with the performance of the corresponding offline modules.

Most commonly used criteria for speaker change detection like log likelihood ratio (LLR) and Bayesian information criterion (BIC) have an adjustable threshold/penalty parameter to make speaker change decisions. These parameters are not always robust to different acoustic conditions and have to be tuned. In this letter, we present a criterion which can be used to identify speaker changes in an audio stream without such tuning. The criterion consists of calculating the LLR of two models with the same number of parameters. Results on the Hub4 1997 evaluation set indicate that we achieve a performance comparable to using BIC with optimal penalty term.

The paper proposes a technique that segments audio according to speakers and based on their location. In many multi-party conversations, such as meetings, the location of participants is restricted to a small number of regions, such as seats around a table, or at a whiteboard. In such cases, segmentation according to these discrete regions would be a reliable means of determining speaker turns. We propose a system that uses microphone pair time delays as features to represent speaker locations. These features are integrated in a GMM/HMM framework to determine an optimal segmentation of the audio according to location. The HMM framework also allows extensions to recognise more complex structures, such as the presence of two simultaneous speakers. Experiments testing the system on real recordings from a meeting room show that the proposed location features can provide greater discrimination than standard cepstral features, and also demonstrate the success of an extension to handle dual-speaker overlap.

The paper investigates the recognition of group actions in meetings by modeling the joint behaviour of participants. Many meeting actions, such as presentations, discussions and consensus, are characterised by similar or complementary behaviour across participants. Recognising these meaningful actions is an important step towards the goal of providing effective browsing and summarisation of processed meetings. A corpus of meetings was collected in a room equipped with a number of microphones and cameras. The corpus was labeled in terms of a predefined set of meeting actions characterised by global behaviour. In experiments, audio and visual features for each participant are extracted from the raw data and the interaction of participants is modeled using HMM-based approaches. Initial results on the corpus demonstrate the ability of the system to recognise the set of meeting actions.

This paper investigates the use of microphone arrays to acquire and recognise speech in meetings. Meetings pose several interesting problems for speech processing, as they consist of multiple competing speakers within a small space, typically around a table. Due to their ability to provide hands-free acquisition and directional discrimination, microphone arrays present a potential alternative to close-talking microphones in such an application. We first propose an appropriate microphone array geometry and improved processing technique for this scenario, paying particular attention to speaker separation. during possible overlap segments. Data collection of a small vocabulary speech recognition corpus (Numbers) was performed in a real meeting room for a single speaker, and several overlapping speech scenarios. In speech recognition experiments on the acquired database, the performance of the microphone array system is compared to that of a close-talking lapel microphone, and a single table-top microphone.

In this paper, we discuss meetings as an application domain for multimedia content analysis. Meeting databases are a rich data source suitable for a variety of audio, visual and multi-modal tasks, including speech recognition, people and action recognition, and information retrieval. We specifically focus on the task of semantic annotation of audio-visual (AV) events, where annotation consists of assigning labels (event names) to the data. In order to develop an automatic annotation system in a principled manner, it is essential to have a well-defined task, a standard corpus and an objective performance measure. In this work we address each of these issues to automatically annotate events based on participant interactions.

We present a probabilistic method for audio-visual (AV) speaker tracking, using an uncalibrated wide-angle camera and a micro- phone array. The algorithm fuses 2-D object shape and audio information via importance particle filters (I-PFs), allowing for the asymmetrical integration of AV information in a way that efficiently exploits the complementary features of each modality. Audio localization information is used to generate an importance sampling (IS) function, which guides the random search process of a particle filter towards regions of the configuration space likely to contain the true configuration (a speaker). The measurement process integrates contour-based and audio observations, which results in reliable head tracking in realistic scenarios. We show that imperfect single modalities can be combined into an algorithm that automatically initializes and tracks a speaker, switches between multiple speakers, tolerates visual clutter, and recovers from total AV object occlusion, in the context of a multimodal meeting room.

This paper proposes a technique that segments audio according to speakers based on their location. In many multi-party conversations, such as meetings, the location of participants is restricted to a small number of regions, such as seats around a table, or at a whiteboard. In such cases, segmentation according to these discrete regions would be a reliable means of determining speaker turns. We propose a system that uses microphone pair time delays as features to represent speaker locations. These features are integrated in a GMM/HMM framework to determine an optimal segmentation of the audio according to location. The HMM framework also allows extensions to recognize more complex structure, such as the presence of two simultaneous speakers. Experiments testing the system on real recordings from a meeting room show that the proposed location features can provide greater discrimination than standard cepstral features, and also demonstrate the success of an extension to handle dual-speaker overlap.

Speaker turn detection is an important task for many speech processing applications. However, accurate segmentation can be hard to achieve if there are multiple concurrent speakers (overlap), as is typically the case in multi-party conversations. In such cases, the location of the speaker, as measured using a microphone array, may provide greater discrimination than traditional spectral features. This was verified in previous work which obtained a global segmentation in terms of single speaker classes, as well as possible overlap combinations. However, such a global strategy suffers from an explosion of the number of overlap classes, as each possible combination of concurrent speakers must be modeled explicitly. In this paper, we propose two alternative schemes that produce an individual segmentation decision for each speaker, implicitly handling all overlapping speaker combinations. The proposed approaches also allow straightforward online implementations. Experiments are presented comparing the segmentation with that obtained using the previous system. 1.

In this paper, we introduce new dynamic speech features based on the modulation spectrum. These features, termed Mel-cepstrum Modulation Spectrum (MCMS), map the time trajectories of the spectral dynamics into a series of slow and fast moving orthogonal components, providing a more general and discriminative range of dynamic features than traditional delta and acceleration features. The features can be seen as the outputs of an array of band-pass filters spread over the cepstral modulation frequency range of interest. In experiments, it is shown that, as well as providing a slight improvement in clean conditions, these new dynamic features yield a significant increase in speech recognition performance in various noise conditions when compared directly to the standard temporal derivative features and RASTA-PLP features.

This paper introduces a novel technique for estimating the signal power spectral density to be used in the transfer function of a microphone array post-filter. The technique is a generalization of the existing Zelinski post-filter, which uses the auto- and cross-spectral densities of the array inputs to estimate the signal and noise spectral densities. The Zelinski technique, however, assumes zero cross-correlation between the noise on different sensors. This assumption is inaccurate, particularly at low frequencies and for arrays with closely spaced sensors, and thus the corresponding post-filter is suboptimal in realistic noise conditions. In this paper, a more general expression of the post-filter estimation is developed based on an assumed knowledge of the complex coherence of the noise field. This general expression can be used to construct a more appropriate post-filter in a variety of different noise fields. In experiments using real noise recordings from a computer office, the modified post-filter results in significant improvement in terms of objective speech quality measures and speech recognition performance using a diffuse noise model.

In this paper we present a new approach towards high performance speech/music segmentation on realistic tasks related to the automatic transcription of broadcast news. In the approach presented here, the local probability density function (PDF) estimators trained on clean microphone speech are used as a channel model at the output of which the entropy and “dynamism” will be measured and integrated over time through a 2-state (speech and and non-speech) hidden Markov model (HMM) with minimum duration constraints. The parameters of the HMM are trained using the EM algorithm in a completely unsupervised manner. Different experiments, including a variety of speech and music styles, as well as different segment durations of speech and music signals (real data distribution, mostly speech, or mostly music), will illustrate the robustness of the approach, which in each case achieves a frame-level accuracy greater than 94%.

This paper proposes a novel technique for estimating the signal power spectral density to be used in the transfer function of a microphone array post-filter. The technique is a modification of the existing Zelinski post-filter, which uses the auto- and cross-spectral densities of the array inputs to estimate the signal and noise spectral densities. The Zelinski technique, however, assumes zero cross-correlation between noise on different sensors. This assumption is inaccurate in real conditions, particularly at low frequencies and for arrays with closely spaced sensors. In this paper we replace this with an assumption of a theoretically diffuse noise field, which is more appropriate in a variety of realistic noise environments. In experiments using noise recordings from an office of computer workstations, the modified post-filter results in significant improvement in terms of objective speech quality measures and speech recognition performance.

An HMM-based speaker clustering framework is presented, where the number of speakers and segmentation boundaries are unknown \\emph{a priori}. Ideally, the system aims to create one pure cluster for each speaker. The HMM is ergodic in nature with a minimum duration topology. The final number of clusters is determined automatically by merging closest clusters and retraining this new cluster, until a decrease in likelihood is observed. In the same framework, we also examine the effect of using only the features from highly voiced frames as a means of improving the robustness and computational complexity of the algorithm. The proposed system is assessed on the 1996 HUB-4 evaluation test set in terms of both cluster and speaker purity. It is shown that the number of clusters found often correspond to the actual number of speakers.

Traditional microphone array speech recognition systems simply recognise the enhanced output of the array. As the level of signal enhancement depends on the number of microphones, such systems do not achieve acceptable speech recognition performance for arrays having only a few microphones. For small microphone arrays, we instead propose using the enhanced output to estimate a reliability mask, which is then used in missing data speech recognition. In missing data speech recognition, the decoded sequence depends on the reliability of each input feature. This reliability is usually based on the signal to noise ratio in each frequency band. In this paper, we use the energy difference between the noisy input and the enhanced output of a small microphone array to determine the frequency band reliability. Recognition experiments with a small array demonstrate the effectiveness of the technique, compared to both traditional microphone array enhancement and a baseline missing data system.

Proposes the integration of sub-band speech recognition with a microphone array. A broadband beamforming microphone array allows for natural integration with sub-band speech recognition as the beamformer is typically implemented as a combination of band-limited sub-arrays. In the paper, rather than recombining the sub-array outputs to give a single enhanced output, we propose the fusion of separate hidden Markov models trained on each subarray frequency band. In addition, a dynamic sub-band weighting scheme is proposed in which the cross- and auto-spectral densities of the microphone array inputs are used to estimate the reliability of each frequency band. The microphone array sub-band system is evaluated on an isolated digit recognition task and compared to the standard full-band approach. The results of the proposed dynamic weighting scheme are compared to those obtained using both fixed equal sub-band weights, as well as optimal sub-band weights calculated from a priori knowledge of the correct results.

This paper assesses the effectiveness of a microphone array system in providing robust hands-free speech recognition on a computer workstation. Near-field superdirectivity (NFSD) is used for the beamforming, and the effect of adding a post-filter is investigated. In the experiments, the speaker is located directly in front of the computer monitor at a distance of 60 cm and the array is designed to fit across the top of a standard 17 inch monitor. To test the proposed system, connected digit strings from the TIDIGITS database are used. The array is shown to have a marked effect on the recognition results in a high noise office environment, particularly where there is a high level of undesired speech.