Acoustic Recognition Research Articles

Design of an Acoustic Target Classification System Based on Small-Aperture Microphone Array

The acoustic recognition module of the unattended ground sensor (UGS) system applied in wild environments is faced with the challenge of complicated noise interference. In this paper, a small-aperture microphone array (MA)-based acoustic target classification system, including the system hardware architecture and classification algorithm scheme, is designed as a node-level sensor for the application of UGS in noisy situation. Starting from the analysis of signature of the acoustic signal in wild environments and the merits of small-aperture array in noise reduction, a closely arranged microelectromechanical systems MA is designed to improve the signal quality. Considering the similarities between speaker discrimination and acoustic target recognition, a classification algorithm scheme, consisting of a simplified Mel-frequency cepstrum coefficients and the Gaussian mixture model, is developed to distinguish acoustic targets’ patterns. The proposed classification algorithm has been implemented on embedded system after being tested on training datasets. By combining the small-aperture array and low-complexity classification algorithm, the presented acoustic classification prototype system is portable and efficient. To demonstrate the efficiency of the design, the prototype system is verified in a practical situation with the employment of wheeled and tracked vehicles. Evaluation of the system performances in comparison with other state-of-the-art methods indicates that the proposed design is practical for the acoustic target classification and may be widely adopted by UGS.

Bird acoustic activity detection based on morphological filtering of the spectrogram

Audio event recognition methods based on the Hidden Markov Model / Gaussian Mixture Model (HMM/GMM) often depend on a large number of mixture components or multi-stage models that require significant computational and memory resources during their operation. A widely used approach for coping with complexity is employing an acoustic activity detector, which selects for further processing only those portions of the audio that are considered promising. As a result, the audio feature extraction and the subsequent pattern recognition stages will process only a subset of the original audio stream, helping to reduce the misclassification rates while lowering the computational demands. In the present work we propose a method for bird acoustic activity detection, based on morphological filtering of the spectrogram seen as an image. The practical significance of the proposed method is validated on the automated acoustic recognition of Southern Lapwing Vanellus chilensis, a common Neotropical bird species. Compared with other methods of acoustic activity detection it demonstrates advantageous performance.

Automated acoustic detection of Vanellus chilensis lampronotus

Automated acoustic recognition of Southern Lapwing in real-world soundscapes.Recognizer of Vanellus chilensis vocalizations and their start and end timestamps.Log-likelihood ratio estimator with temporal post-processing of the output scores.Computer-assisted analysis of daily and hourly acoustic activity patterns.Non-intrusive monitoring of presence/absence and activity patterns. Traditional human-observer-based biological surveys are expensive. Therefore most biodiversity studies are implemented only periodically, for short periods, and predominantly during daytime and under favorable weather conditions. Automated data acquisition and analysis can overcome these shortcomings and facilitate continuous monitoring. Here we report on the development of an automated acoustic recognizer for Southern Lapwing Vanellus chilensis lampronotus vocalizations, a first for this species. The recognizer is a species-specific information retrieval agent, which searches throughout long audio recordings in order to detect and timestamp call events of the target species. The recognizer relies on a log-likelihood ratio estimator, based on a Gaussian Mixture Model-Universal Background Model (GMM-UBM), complemented with purposely-developed temporal post-processing that incorporates domain knowledge about the structure of V. chilensis vocalizations. Validation experiments with real-field recordings of complex soundscapes indicate that the recognizer is sensitive enough to register V. chilensis call events with sound levels down to -30dB and recognition accuracy of up to 85.6%, at zero false positive rates. The recognizer is considered a valuable tool for computer-assisted analysis of hourly and daily acoustic activity of V. chilensis over extended periods of time, as it offers an indispensable support to long-term monitoring studies and conservation efforts in the Pantanal region.1INAU Project 3.14 Monitoring Bioindicators and Migratory Birds in the Pantanal. Applied Acoustomics - a Tool for Bio-sustainability Assessment, INAU Lab. 3 - Biodiversity and Ecological Processes (INAU: www.inau.org.br; 2011-2014) - Program CsF (www.cienciasemfronteiras.gov.br/web/csf).1

Butterflies that trick ants with sound

Larvae of parasitic Maculinea butterflies use acoustic mimicry to induce host ants to adopt and care for them.

A Robust Feature Extraction Algorithm for the Classification of Acoustic Targets in Wild Environments

The acoustic recognition technology of unattended ground sensor systems applied in wild environments is faced with the challenge of complicated and strong acoustic noise, especially wind noise. Moreover, the commonly used Mel-frequency cepstral coefficients (MFCCs) are sensitive to noise interference. To resolve the problem, a robust feature extraction method, called harmonic Mel-frequency cepstral coefficients (HMFCCs), is proposed for acoustic target classification. By combining an acoustic signal's harmonic model with the MFCC method, the HMFCC has the ability to emphasize the signals emitted by the principal acoustic components of the target. In the experiment conducted for this study, three data sets are sampled under the same conditions, except for wind power levels. Then the classifier, which is trained by one of the three data sets, is used to recognize the others data sets. According to the experimental results, the HMFCC-based classification accuracies of the three data sets are higher than those of other state-of-the-art methods, indicating that HMFCC is a kind of noise-insensitive feature.

An Acoustic Events Recognition for Robotic Systems Based on a Deep Learning Method

In this paper, we provide a new approach to classify and recognize the acoustic events for multiple autonomous robots systems based on the deep learning mechanisms. For disaster response robotic systems, recognizing certain acoustic events in the noisy environment is very effective to perform a given operation. As a new approach, trained deep learning networks which are constructed by RBMs, classify the acoustic events from input waveform signals. From the experimental results, usefulness of our approach is discussed and verified.

Acoustic recognition of multiple bird species based on penalised maximum likelihood

Automatic system for recognition of multiple bird species in audio recordings is presented. Time-frequency segmentation of the acoustic scene is obtained by employing a sinusoidal detection algorithm, which does not require any estimate of noise and is able to handle multiple simultaneous bird vocalizations. Each segment is characterized as a sequence of frequencies over time, referred to as a frequency track. Each bird species is represented by a hidden Markov model that models the temporal evolution of frequency tracks. The decision on the number and identity of bird species in a given recording is obtained based on maximizing the overall likelihood of the set of detected segments, with a penalization applied for increasing the number of bird models used. Experimental evaluations are performed on audio field recordings containing 30 bird species. The presence of multiple bird species is simulated by joining the set of detected segments from several bird species. Results show that the proposed method can achieve recognition performance for multiple bird species not far from that obtained for single bird species, and considerably outperforms majority voting methods.

A Hidden Markov Model-Based Acoustic Cicada Detector for Crowdsourced Smartphone Biodiversity Monitoring

In recent years, the field of computational sustainability has striven to apply artificial intelligence techniques to solve ecological and environmental problems. In ecology, a key issue for the safeguarding of our planet is the monitoring of biodiversity. Automated acoustic recognition of species aims to provide a cost-effective method for biodiversity monitoring. This is particularly appealing for detecting endangered animals with a distinctive call, such as the New Forest cicada. To this end, we pursue a crowdsourcing approach, whereby the millions of visitors to the New Forest, where this insect was historically found, will help to monitor its presence by means of a smartphone app that can detect its mating call. Existing research in the field of acoustic insect detection has typically focused upon the classification of recordings collected from fixed field microphones. Such approaches segment a lengthy audio recording into individual segments of insect activity, which are independently classified using cepstral coefficients extracted from the recording as features. This paper reports on a contrasting approach, whereby we use crowdsourcing to collect recordings via a smartphone app, and present an immediate feedback to the users as to whether an insect has been found. Our classification approach does not remove silent parts of the recording via segmentation, but instead uses the temporal patterns throughout each recording to classify the insects present. We show that our approach can successfully discriminate between the call of the New Forest cicada and similar insects found in the New Forest, and is robust to common types of environment noise. A large scale trial deployment of our smartphone app collected over 6000 reports of insect activity from over 1000 users. Despite the cicada not having been rediscovered in the New Forest, the effectiveness of this approach was confirmed for both the detection algorithm, which successfully identified the same cicada through the app in countries where the same species is still present, and of the crowdsourcing methodology, which collected a vast number of recordings and involved thousands of contributors.

A Generic and Scalable Architecture for a Large Acoustic Model and Large Vocabulary Speech Recognition Accelerator Using Logic on Memory

This paper describes a scalable hardware accelerator for speech recognition, which uses a two pass decoding algorithm with word dependent N-best Viterbi Beam Search. The observation probability calculation (Senone scoring) and first pass of decoding using a Bigram language model is implemented in hardware. The word lattice output from the first pass is used by software for the second pass, with a trigram language model. The proposed design uses a logic-on-memory approach to make use of high bandwidth nor flash memory to improve random read performance for Senone scoring and first pass decoding, both of which are memory intensive operations. The proposed HW/SW co-design achieves an overall speed up of 4.3X over a 2.4-GHz Intel Core 2 Duo processor running the CMU Sphinx speech recognition software, while consuming an estimated 1.72 W of power. The hardware accelerator provides improved speech recognition accuracy by supporting larger acoustic models and word dictionaries while maintaining real-time performance.

Homogenous ensemble phonotactic language recognition based on SVM supervector reconstruction

Abstract Currently, acoustic spoken language recognition (SLR) and phonotactic SLR systems are widely used language recognition systems. To achieve better performance, researchers combine multiple subsystems with the results often much better than a single SLR system. Phonotactic SLR subsystems may vary in the acoustic features vectors or include multiple language-specific phone recognizers and different acoustic models. These methods achieve good performance but usually compute at high computational cost. In this paper, a new diversification for phonotactic language recognition systems is proposed using vector space models by support vector machine (SVM) supervector reconstruction (SSR). In this architecture, the subsystems share the same feature extraction, decoding, and N-gram counting preprocessing steps, but model in a different vector space by using the SSR algorithm without significant additional computation. We term this a homogeneous ensemble phonotactic language recognition (HEPLR) system. The system integrates three different SVM supervector reconstruction algorithms, including relative SVM supervector reconstruction, functional SVM supervector reconstruction, and perturbing SVM supervector reconstruction. All of the algorithms are incorporated using a linear discriminant analysis-maximum mutual information (LDA-MMI) backend for improving language recognition evaluation (LRE) accuracy. Evaluated on the National Institute of Standards and Technology (NIST) LRE 2009 task, the proposed HEPLR system achieves better performance than a baseline phone recognition-vector space modeling (PR-VSM) system with minimal extra computational cost. The performance of the HEPLR system yields 1.39%, 3.63%, and 14.79% equal error rate (EER), representing 6.06%, 10.15%, and 10.53% relative improvements over the baseline system, respectively, for the 30-, 10-, and 3-s test conditions.

Context-based environmental audio event recognition for scene understanding

Automatic audio content recognition has attracted an increasing attention for developing multimedia systems, for which the most popular approaches combine frame-based features with statistic models or discriminative classifiers. The existing methods are effective for clean single-source event detection but may not perform well for unstructured environmental sounds, which have a broad noise-like flat spectrum and a diverse variety of compositions. We present an automatic acoustic scene understanding framework that detects audio events through two hierarchies, acoustic scene recognition and audio event recognition, in which the former is preceded by following dominant audio sources and in turn helps infer non-dominant audio events within the same scene through modeling their occurrence correlations. On the scene recognition hierarchy, we perform adaptive segmentation and feature extraction for every input acoustic scene stream through Eigen-audiospace and an optimized feature subspace, respectively. After filtering background, scene streams are recognized by modeling the observation density of dominant features using a two-level hidden Markov model. On the audio event recognition hierarchy, scene knowledge is characterized by an audio context model that essentially describes the occurrence correlations of dominant and non-dominant audio events within this scene. Monte Carlo integration and gradient descent techniques are employed to maximize the likelihood and correctly tag each audio event. To the best of our knowledge, this is the first work that models event correlations as scene context for robust audio event detection from complex and noisy environments. Note that according to the recent report, the mean accuracy for the acoustic scene classification task by human listeners is only around 71 % on the data collected in office environments from the DCASE dataset. None of the existing methods performs well on all scene categories and the average accuracy of the best performances of the recent 11 methods is 53.8 %. The proposed method averagely achieves an accuracy of 62.3 % on the same dataset. Additionally, we create a 10-CASE dataset by manually collecting 5,250 audio clips of 10 scene types and 21 event categories. Our experimental results on 10-CASE show that the proposed method averagely achieves the enhanced performance of 78.3 %, and the average accuracy of audio event recognition can be effectively improved by capturing dominant audio sources and reasoning non-dominant events from the dominant ones through acoustic context modeling. In the future work, exploring the interactions between acoustic scene recognition and audio event detection, and incorporating other modalities to improve the accuracy are required to further advance the proposed framework.

Discriminative Minimum Statistics Projection Coefficient Feature for Acoustic Context Recognition

Acoustic environment recognition, which can provide the important acoustic context, has been widely used in many applications and is a considerable difficult problem in the real-life and the complex environment. This paper proposes the discriminative minimum statistics project coefficient (MSPC) feature with the information of classification by using partial least squares (PLS). With the minimum statistics (MS) tracked from the input sound, the discriminative MSPC feature is extracted by projecting the MS into lower-dimensional feature subspace learned by using PLS analysis. Based on the proposed discriminative MSPC feature, the acoustic environment recognition is implemented by using Gaussian Mixture Model (GMM) for modeling each sound class. The experimental results show that the proposed discriminative MSPC feature based on PLS outperforms the MSPC feature based on PCA for acoustic environment recognition.

Mismatched feature detection with finer granularity for emotional speaker recognition

The shapes of speakers’ vocal organs change under their different emotional states, which leads to the deviation of the emotional acoustic space of short-time features from the neutral acoustic space and thereby the degradation of the speaker recognition performance. Features deviating greatly from the neutral acoustic space are considered as mismatched features, and they negatively affect speaker recognition systems. Emotion variation produces different feature deformations for different phonemes, so it is reasonable to build a finer model to detect mismatched features under each phoneme. However, given the difficulty of phoneme recognition, three sorts of acoustic class recognition—phoneme classes, Gaussian mixture model (GMM) tokenizer, and probabilistic GMM tokenizer—are proposed to replace phoneme recognition. We propose feature pruning and feature regulation methods to process the mismatched features to improve speaker recognition performance. As for the feature regulation method, a strategy of maximizing the between-class distance and minimizing the within-class distance is adopted to train the transformation matrix to regulate the mismatched features. Experiments conducted on the Mandarin affective speech corpus (MASC) show that our feature pruning and feature regulation methods increase the identification rate (IR) by 3.64% and 6.77%, compared with the baseline GMM-UBM (universal background model) algorithm. Also, corresponding IR increases of 2.09% and 3.32% can be obtained with our methods when applied to the state-of-the-art algorithm i-vector.

Time and timing in the acoustic recognition system of crickets.

The songs of many insects exhibit precise timing as the result of repetitive and stereotyped subunits on several time scales. As these signals encode the identity of a species, time and timing are important for the recognition system that analyzes these signals. Crickets are a prominent example as their songs are built from sound pulses that are broadcast in a long trill or as a chirped song. This pattern appears to be analyzed on two timescales, short and long. Recent evidence suggests that song recognition in crickets relies on two computations with respect to time; a short linear-nonlinear (LN) model that operates as a filter for pulse rate and a longer integration time window for monitoring song energy over time. Therefore, there is a twofold role for timing. A filter for pulse rate shows differentiating properties for which the specific timing of excitation and inhibition is important. For an integrator, however, the duration of the time window is more important than the precise timing of events. Here, we first review evidence for the role of LN-models and integration time windows for song recognition in crickets. We then parameterize the filter part by Gabor functions and explore the effects of duration, frequency, phase, and offset as these will correspond to differently timed patterns of excitation and inhibition. These filter properties were compared with known preference functions of crickets and katydids. In a comparative approach, the power for song discrimination by LN-models was tested with the songs of over 100 cricket species. It is demonstrated how the acoustic signals of crickets occupy a simple 2-dimensional space for song recognition that arises from timing, described by a Gabor function, and time, the integration window. Finally, we discuss the evolution of recognition systems in insects based on simple sensory computations.

Gorgon Project: environmental acoustic recognition sensor (EARS)

The Asian house gecko (Hemidactylus frenatus) presents a significant risk to the unique environment of Barrow Island, the location of the Chevron-operated Gorgon Project. The island’s climate would suit the establishment of this non-indigenous species, which could displace the five known native gecko species. Originating in Asia, this pest has spread to almost 9 new locations by hitchhiking on international vessels and cargo. To help detect possible gecko activity on the Gorgon Project, a unique device has been developed to specifically detect the sound of the gecko call. EARS can detect the sound of a gecko’s call from up to 2 m away. Connected to a wireless network, it sends an alert when a call is registered detailing time, location, and a sound file to aid detection. Placed on cargo and vessels travelling to Barrow, it is also deployed in Australian supply bases, international fabrication yards, and marine offloading facilities on the island. EARS provide Gorgon Quarantine with a constant presence and are particularly useful for detecting nocturnal species. It complements Gorgon’s existing quarantine surveillance programs and is one of many measures to ensure Barrow Island remains free from non-indigenous species. The Gorgon Project is operated by an Australian subsidiary of Chevron and is a joint venture of the Australian subsidiaries of Chevron (47.3%), ExxonMobil (25%), Shell (25%), Osaka Gas (1.25%), Tokyo Gas (1%) and Chubu Electric Power (0.417%).

Coal Gangue Interface Recognition Based on MFCC Research

In view of the existing coal gangue interface identification technology use ray method that it is not applicable to the working face what it dose not contain radioactive elements in the roof or contain low amounts of radioactive elements, and the detection range of radar detection is lesser, the signal attenuation is relatively serious, so putting forward a kind of coal gangue interface recognition based on Mel frequency cepstrum coefficient of MFCC. The method using coal gangue were put down in the process of coal gangue on the difference of the characteristics of the acoustic signal recognition, first of all, using Mel frequency cepstrum coefficient will after denoising coal gangue acoustic signal transformation to the frequency domain processing, extracting the 32 d characteristic parameters of coal and gangue acoustic signal;The experimental results show that the method can accurately identify the coal gangue down state.

Visual Features for Lithuanian Phone Recognition

Paper presents visual features for speech recognition. Visual speech recognition can be applied as support for acoustical recognition process or as stand-alone speech recognition approach in the case of absence of acoustical data. Geometrical features of speaker‘s lips are presented. Statistical analysis of feature values enabled us to select individual feature for each analyzed phone and one combined feature for all phones. The discriminative power was tested using Bayes classification rule. 4th order analysis enabled us to recognize phones with the average rate of 60.3 %, the highest recognition rate of 81.3 % was achieved identifying phone [o]. DOI: http://dx.doi.org/10.5755/j01.eee.19.5.1724

Familiarity perception call elicited under restricted sensory cues in peer-social interactions of the domestic chick.

Social cognitive mechanisms are central to understanding developmental abnormalities, such as autistic spectrum disorder. Peer relations besides parent-infant or pair-bonding interactions are pivotal social relationships that are especially well developed in humans. Cognition of familiarity forms the basis of peer socialization. Domestic chick (Gallus gallus) studies have contributed to our understanding of the developmental process in sensory-motor cognition but many processes remain unknown. In this report, we used chicks, as they are precocial birds, and we could therefore focus on peer interaction without having to consider parenting. The subject chick behavior towards familiar and unfamiliar reference peers was video-recorded, where the subject and the reference were separated by either an opaque or transparent wall. Spectrogram and behavior correlation analyses based on principal component analysis, revealed that chicks elicited an intermediate contact call and a morphologically different distress call, more frequently towards familiar versus unfamiliar chicks in acoustic only conditions. When both visual and acoustic cues were present, subject chicks exhibited approaching and floor pecking behavior, while eliciting joyful (pleasant) calls, irrespective of whether reference peers were familiar or unfamiliar. Our result showed that chicks recognized familiarity using acoustic cues and expressed cognition through modified distress calls. These finding suggests that peer affiliation may be established by acoustic recognition, independent of visual face recognition, and that eventually, both forms of recognition are integrated, with modulation of acoustic recognition.

Interspecific acoustic recognition in two European bat communities

Echolocating bats emit echolocation calls for spatial orientation and foraging. These calls are often species-specific and are emitted at high intensity and repetition rate. Therefore, these calls could potentially function in intra- and/or inter-specific bat communication. For example, bats in the field approach playbacks of conspecific feeding buzzes, probably because feeding buzzes indicate an available foraging patch. In captivity, some species of bats recognize and distinguish the echolocation calls of different sympatric species. However, it is still unknown if and how acoustic species-recognition mediates interspecific interactions in the field. Here we aim to understand eavesdropping on bat echolocation calls within and across species boundaries in wild bats. We presented playbacks of conspecific and heterospecific search calls and feeding buzzes to four bat species with different foraging ecologies. The bats were generally more attracted by feeding buzzes than search calls and more by the calls of conspecifics than their heterospecifics. Furthermore, bats showed differential reaction to the calls of the heterospecifics. In particular, Myotis capaccinii reacted equally to the feeding buzzes of conspecifics and to ecologically more similar heterospecifics. Our results confirm eavesdropping on feeding buzzes at the intraspecific level in wild bats and provide the first experimental quantification of potential eavesdropping in European bats at the interspecific level. Our data support the hypothesis that bat echolocation calls have a communicative potential that allows interspecific, and potentially intraspecific, eavesdropping in the wild.

Sensory Ecology of Water Detection by Bats: A Field Experiment

Bats face a great risk of dehydration, so sensory mechanisms for water recognition are crucial for their survival. In the laboratory, bats recognized any smooth horizontal surface as water because these provide analogous reflections of echolocation calls. We tested whether bats also approach smooth horizontal surfaces other than water to drink in nature by partly covering watering troughs used by hundreds of bats with a Perspex layer mimicking water. We aimed 1) to confirm that under natural conditions too bats mistake any horizontal smooth surface for water by testing this on large numbers of individuals from a range of species and 2) to assess the occurrence of learning effects. Eleven bat species mistook Perspex for water relying chiefly on echoacoustic information. Using black instead of transparent Perspex did not deter bats from attempting to drink. In Barbastella barbastellus no echolocation differences occurred between bats approaching the water and the Perspex surfaces respectively, confirming that bats perceive water and Perspex to be acoustically similar. The drinking attempt rates at the fake surface were often lower than those recorded in the laboratory: bats then either left the site or moved to the control water surface. This suggests that bats modified their behaviour as soon as the lack of drinking reward had overridden the influence of echoacoustic information. Regardless of which of two adjoining surfaces was covered, bats preferentially approached and attempted to drink from the first surface encountered, probably because they followed a common route, involving spatial memory and perhaps social coordination. Overall, although acoustic recognition itself is stereotyped and its importance in the drinking process overwhelming, our findings point at the role of experience in increasing behavioural flexibility under natural conditions.