When watching videos, the occurrence of a visual event is often accompanied by an audio event, e.g., the voice of lip motion, the music of playing instruments. There is an underlying correlation between audio and visual events, which can be utilized as free supervised information to train a neural network by solving the pretext task of audio-visual synchronization. In this paper, we propose a novel self-supervised framework with co-attention mechanism to learn generic cross-modal representations from unlabelled videos in the wild, and further benefit downstream tasks. Specifically, we explore three different co-attention modules to focus on discriminative visual regions correlated to the sounds and introduce the interactions between them. Experiments show that our model achieves state-of-the-art performance on the pretext task while having fewer parameters compared with existing methods. To further evaluate the generalizability and transferability of our approach, we apply the pre-trained model on two downstream tasks, i.e., sound source localization and action recognition. Extensive experiments demonstrate that our model provides competitive results with other self-supervised methods, and also indicate that our approach can tackle the challenging scenes which contain multiple sound sources.
Detecting sound source objects within visual observation is important for autonomous robots to comprehend surrounding environments. Since sounding objects have a large variety with different appearances in our living environments, labeling all sounding objects is impossible in practice. This calls for self-supervised learning which does not require manual labeling. Most of conventional self-supervised learning uses monaural audio signals and images and cannot distinguish sound source objects having similar appearances due to poor spatial information in audio signals. To solve this problem, this paper presents a self-supervised training method using 360{\deg} images and multichannel audio signals. By incorporating with the spatial information in multichannel audio signals, our method trains deep neural networks (DNNs) to distinguish multiple sound source objects. Our system for localizing sound source objects in the image is composed of audio and visual DNNs. The visual DNN is trained to localize sound source candidates within an input image. The audio DNN verifies whether each candidate actually produces sound or not. These DNNs are jointly trained in a self-supervised manner based on a probabilistic spatial audio model. Experimental results with simulated data showed that the DNNs trained by our method localized multiple speakers. We also demonstrate that the visual DNN detected objects including talking visitors and specific exhibits from real data recorded in a science museum.
During the performance of sound source localization which uses both visual and aural information, it presently remains unclear how much either image or sound modalities contribute to the result, i.e. do we need both image and sound for sound source localization? To address this question, we develop an unsupervised learning system that solves sound source localization by decomposing this task into two steps: (i) "potential sound source localization", a step that localizes possible sound sources using only visual information (ii) "object selection", a step that identifies which objects are actually sounding using aural information. Our overall system achieves state-of-the-art performance in sound source localization, and more importantly, we find that despite the constraint on available information, the results of (i) achieve similar performance. From this observation and further experiments, we show that visual information is dominant in "sound" source localization when evaluated with the currently adopted benchmark dataset. Moreover, we show that the majority of sound-producing objects within the samples in this dataset can be inherently identified using only visual information, and thus that the dataset is inadequate to evaluate a system's capability to leverage aural information. As an alternative, we present an evaluation protocol that enforces both visual and aural information to be leveraged, and verify this property through several experiments.
Inspired by the mammal's auditory localization pathway, in this paper we propose a pure spiking neural network (SNN) based computational model for precise sound localization in the noisy real-world environment, and implement this algorithm in a real-time robotic system with a microphone array. The key of this model relies on the MTPC scheme, which encodes the interaural time difference (ITD) cues into spike patterns. This scheme naturally follows the functional structures of the human auditory localization system, rather than artificially computing of time difference of arrival. Besides, it highlights the advantages of SNN, such as event-driven and power efficiency. The MTPC is pipelined with two different SNN architectures, the convolutional SNN and recurrent SNN, by which it shows the applicability to various SNNs. This proposal is evaluated by the microphone collected location-dependent acoustic data, in a real-world environment with noise, obstruction, reflection, or other affects. The experiment results show a mean error azimuth of 1~3 degrees, which surpasses the accuracy of the other biologically plausible neuromorphic approach for sound source localization.
Telepresence robots are used in various forms in various use-cases that helps to avoid physical human presence at the scene of action. In this work, we focus on a telepresence robot that can be used to attend a meeting remotely with a group of people. Unlike a one-to-one meeting, participants in a group meeting can be located at a different part of the room, especially in an informal setup. As a result, all of them may not be at the viewing angle of the robot, a.k.a. the remote participant. In such a case, to provide a better meeting experience, the robot should localize the speaker and bring the speaker at the center of the viewing angle. Though sound source localization can easily be done using a microphone-array, bringing the speaker or set of speakers at the viewing angle is not a trivial task. First of all, the robot should react only to a human voice, but not to the random noises. Secondly, if there are multiple speakers, to whom the robot should face or should it rotate continuously with every new speaker? Lastly, most robotic platforms are resource-constrained and to achieve a real-time response, i.e., avoiding network delay, all the algorithms should be implemented within the robot itself. This article presents a study and implementation of an attention shifting scheme in a telepresence meeting scenario which best suits the needs and expectations of the collocated and remote attendees. We define a policy to decide when a robot should rotate and how much based on real-time speaker localization. Using user satisfaction study, we show the efficacy and usability of our system in the meeting scenario. Moreover, our system can be easily adapted to other scenarios where multiple people are located.
The ability to localize and track acoustic events is a fundamental prerequisite for equipping machines with the ability to be aware of and engage with humans in their surrounding environment. However, in realistic scenarios, audio signals are adversely affected by reverberation, noise, interference, and periods of speech inactivity. In dynamic scenarios, where the sources and microphone platforms may be moving, the signals are additionally affected by variations in the source-sensor geometries. In practice, approaches to sound source localization and tracking are often impeded by missing estimates of active sources, estimation errors, as well as false estimates. The aim of the LOCAlization and TrAcking (LOCATA) Challenge is an open-access framework for the objective evaluation and benchmarking of broad classes of algorithms for sound source localization and tracking. This article provides a review of relevant localization and tracking algorithms and, within the context of the existing literature, a detailed evaluation and dissemination of the LOCATA submissions. The evaluation highlights achievements in the field, open challenges, and identifies potential future directions.
This paper describes a system that generates speaker-annotated transcripts of meetings by using a microphone array and a 360-degree camera. The hallmark of the system is its ability to handle overlapped speech, which has been an unsolved problem in realistic settings for over a decade. We show that this problem can be addressed by using a continuous speech separation approach. In addition, we describe an online audio-visual speaker diarization method that leverages face tracking and identification, sound source localization, speaker identification, and, if available, prior speaker information for robustness to various real world challenges. All components are integrated in a meeting transcription framework called SRD, which stands for "separate, recognize, and diarize". Experimental results using recordings of natural meetings involving up to 11 attendees are reported. The continuous speech separation improves a word error rate (WER) by 16.1% compared with a highly tuned beamformer. When a complete list of meeting attendees is available, the discrepancy between WER and speaker-attributed WER is only 1.0%, indicating accurate word-to-speaker association. This increases marginally to 1.6% when 50% of the attendees are unknown to the system.
Visual events are usually accompanied by sounds in our daily lives. However, can the machines learn to correlate the visual scene and sound, as well as localize the sound source only by observing them like humans? To investigate its empirical learnability, in this work we first present a novel unsupervised algorithm to address the problem of localizing sound sources in visual scenes. In order to achieve this goal, a two-stream network structure which handles each modality with attention mechanism is developed for sound source localization. The network naturally reveals the localized response in the scene without human annotation. In addition, a new sound source dataset is developed for performance evaluation. Nevertheless, our empirical evaluation shows that the unsupervised method generates false conclusions in some cases. Thereby, we show that this false conclusion cannot be fixed without human prior knowledge due to the well-known correlation and causality mismatch misconception. To fix this issue, we extend our network to the supervised and semi-supervised network settings via a simple modification due to the general architecture of our two-stream network. We show that the false conclusions can be effectively corrected even with a small amount of supervision, i.e., semi-supervised setup. Furthermore, we present the versatility of the learned audio and visual embeddings on the cross-modal content alignment and we extend this proposed algorithm to a new application, sound saliency based automatic camera view panning in 360-degree{\deg} videos.
Multi-modal learning, particularly among imaging and linguistic modalities, has made amazing strides in many high-level fundamental visual understanding problems, ranging from language grounding to dense event captioning. However, much of the research has been limited to approaches that either do not take audio corresponding to video into account at all, or those that model the audio-visual correlations in service of sound or sound source localization. In this paper, we present the evidence, that audio signals can carry surprising amount of information when it comes to high-level visual-lingual tasks. Specifically, we focus on the problem of weakly-supervised dense event captioning in videos and show that audio on its own can nearly rival performance of a state-of-the-art visual model and, combined with video, can improve on the state-of-the-art performance. Extensive experiments on the ActivityNet Captions dataset show that our proposed multi-modal approach outperforms state-of-the-art unimodal methods, as well as validate specific feature representation and architecture design choices.
This paper proposes a real-time system integrating an acoustic material estimation from visual appearance and an on-the-fly mapping in the 3-dimension. The proposed method estimates the acoustic materials of surroundings in indoor scenes and incorporates them to a 3-D occupancy map, as a robot moves around the environment. To estimate the acoustic material from the visual cue, we apply the state-of-the-art semantic segmentation CNN network based on the assumption that the visual appearance and the acoustic materials have a strong association. Furthermore, we introduce an update policy to handle the material estimations during the online mapping process. As a result, our environment map with acoustic material can be used for sound-related robotics applications, such as sound source localization taking into account various acoustic propagation (e.g., reflection).
Inspired by the behavior of humans talking in noisy environments, we propose an embodied embedded cognition approach to improve automatic speech recognition (ASR) systems for robots in challenging environments, such as with ego noise, using binaural sound source localization (SSL). The approach is verified by measuring the impact of SSL with a humanoid robot head on the performance of an ASR system. More specifically, a robot orients itself toward the angle where the signal-to-noise ratio (SNR) of speech is maximized for one microphone before doing an ASR task. First, a spiking neural network inspired by the midbrain auditory system based on our previous work is applied to calculate the sound signal angle. Then, a feedforward neural network is used to handle high levels of ego noise and reverberation in the signal. Finally, the sound signal is fed into an ASR system. For ASR, we use a system developed by our group and compare its performance with and without the support from SSL. We test our SSL and ASR systems on two humanoid platforms with different structural and material properties. With our approach we halve the sentence error rate with respect to the common downmixing of both channels. Surprisingly, the ASR performance is more than two times better when the angle between the humanoid head and the sound source allows sound waves to be reflected most intensely from the pinna to the ear microphone, rather than when sound waves arrive perpendicularly to the membrane.
From a machine learning perspective, the human ability localize sounds can be modeled as a non-parametric and non-linear regression problem between binaural spectral features of sound received at the ears (input) and their sound-source directions (output). The input features can be summarized in terms of the individual's head-related transfer functions (HRTFs) which measure the spectral response between the listener's eardrum and an external point in $3$D. Based on these viewpoints, two related problems are considered: how can one achieve an optimal sampling of measurements for training sound-source localization (SSL) models, and how can SSL models be used to infer the subject's HRTFs in listening tests. First, we develop a class of binaural SSL models based on Gaussian process regression and solve a \emph{forward selection} problem that finds a subset of input-output samples that best generalize to all SSL directions. Second, we use an \emph{active-learning} approach that updates an online SSL model for inferring the subject's SSL errors via headphones and a graphical user interface. Experiments show that only a small fraction of HRTFs are required for $5^{\circ}$ localization accuracy and that the learned HRTFs are localized closer to their intended directions than non-individualized HRTFs.
This paper introduces a new paradigm for sound source lo-calization referred to as virtual acoustic space traveling (VAST) and presents a first dataset designed for this purpose. Existing sound source localization methods are either based on an approximate physical model (physics-driven) or on a specific-purpose calibration set (data-driven). With VAST, the idea is to learn a mapping from audio features to desired audio properties using a massive dataset of simulated room impulse responses. This virtual dataset is designed to be maximally representative of the potential audio scenes that the considered system may be evolving in, while remaining reasonably compact. We show that virtually-learned mappings on this dataset generalize to real data, overcoming some intrinsic limitations of traditional binaural sound localization methods based on time differences of arrival.
The propagation of sound in a shallow water environment is characterized by boundary reflections from the sea surface and sea floor. These reflections result in multiple (indirect) sound propagation paths, which can degrade the performance of passive sound source localization methods. This paper proposes the use of convolutional neural networks (CNNs) for the localization of sources of broadband acoustic radiated noise (such as motor vessels) in shallow water multipath environments. It is shown that CNNs operating on cepstrogram and generalized cross-correlogram inputs are able to more reliably estimate the instantaneous range and bearing of transiting motor vessels when the source localization performance of conventional passive ranging methods is degraded. The ensuing improvement in source localization performance is demonstrated using real data collected during an at-sea experiment.
It is known that adverse environments such as high reverberation and low signal-to-noise ratio (SNR) pose a great challenge to indoor sound source localization. To address this challenge, in this paper, we propose a sound source localization algorithm based on probabilistic neural network, namely Generalized cross correlation Classification Algorithm (GCA). Experimental results for adverse environments with high reverberation time T60 up to 600ms and low SNR such as -10dB show that, the average azimuth angle error and elevation angle error by GCA are only 4.6 degrees and 3.1 degrees respectively. Compared with three recently published algorithms, GCA has increased the success rate on direction of arrival estimation significantly with good robustness to environmental changes. These results show that the proposed GCA can localize accurately and robustly for diverse indoor applications where the site acoustic features can be studied prior to the localization stage.
We propose to use neural networks for simultaneous detection and localization of multiple sound sources in human-robot interaction. In contrast to conventional signal processing techniques, neural network-based sound source localization methods require fewer strong assumptions about the environment. Previous neural network-based methods have been focusing on localizing a single sound source, which do not extend to multiple sources in terms of detection and localization. In this paper, we thus propose a likelihood-based encoding of the network output, which naturally allows the detection of an arbitrary number of sources. In addition, we investigate the use of sub-band cross-correlation information as features for better localization in sound mixtures, as well as three different network architectures based on different motivations. Experiments on real data recorded from a robot show that our proposed methods significantly outperform the popular spatial spectrum-based approaches.
Visual events are usually accompanied by sounds in our daily lives. We pose the question: Can the machine learn the correspondence between visual scene and the sound, and localize the sound source only by observing sound and visual scene pairs like human? In this paper, we propose a novel unsupervised algorithm to address the problem of localizing the sound source in visual scenes. A two-stream network structure which handles each modality, with attention mechanism is developed for sound source localization. Moreover, although our network is formulated within the unsupervised learning framework, it can be extended to a unified architecture with a simple modification for the supervised and semi-supervised learning settings as well. Meanwhile, a new sound source dataset is developed for performance evaluation. Our empirical evaluation shows that the unsupervised method eventually go through false conclusion in some cases. We show that even with a few supervision, false conclusion is able to be corrected and the source of sound in a visual scene can be localized effectively.
The thud of a bouncing ball, the onset of speech as lips open -- when visual and audio events occur together, it suggests that there might be a common, underlying event that produced both signals. In this paper, we argue that the visual and audio components of a video signal should be modeled jointly using a fused multisensory representation. We propose to learn such a representation in a self-supervised way, by training a neural network to predict whether video frames and audio are temporally aligned. We use this learned representation for three applications: (a) sound source localization, i.e. visualizing the source of sound in a video; (b) audio-visual action recognition; and (c) on/off-screen audio source separation, e.g. removing the off-screen translator's voice from a foreign official's speech. Code, models, and video results are available on our webpage: http://andrewowens.com/multisensory