An Overview of Audio Event Detection Methods from Feature Extraction to Classification

Figures & data

Table 1. Audio classifiers in AED.

Methods	Type of classifier	References
Unsupervised Learning Algorithms	Hierarchical and Partition Clustering	(Pomponi and Vinogradov 2013), (Tsunoo et al. 2011), (Park 2009), (Lefèvre and Vincent 2011) (Yang et al. 2013)
	Gaussian Mixture Models (GMM)	(Chuan 2013), (Choi and Chang 2012), (P. Dhanalakshmi et al. 2011b), (Cheng, Sun, and Ji 2010), (Chung-Hsien and Chia-Hsin 2006)
	Hidden Markov Models (HMM)	(Navarathna et al. 2013), (Niessen, Van Kasteren, and Merentitis 2013), (Itoh, Takiguchi, and Ariki 2013), (Wang and Zhang 2012), (Milone et al. 2012), (Ya-Ti et al. 2009), (Scheme, Hudgins, and Parker 2007)
	Neural networks(Self-organizing map, ART)	(P. Dhanalakshmi et al. 2011a), (Schroeder et al. 2011), (Charalampidis, Georgiopoulos, and Kasparis 2000)
Supervised Learning Algorithms	Instance-based or K-nearest-neighbors (KNN)	(Khunarsal, Lursinsap, and Raicharoen 2013), (Ravan and Beheshti 2011), (Liu and Zhang 2012), (Lie, Hong-Jiang, and Hao 2002), (Huang et al. 2009), (Malhotra, Nikolaidis, and Harms 2008)
	Neural Networks (RBF, MLP)	(Balochian, Seidabad, and Rad 2013), (Ganapathy, Rajan, and Hermansky 2011), (Mitra and Wang 2008), (Kotti et al. 2007), (Shen, Shepherd, and Ngu 2006), (Turnbull and Elkan 2005), (Khairnar, Merchant, and Desai 2005), (McConaghy et al. 2003)
	Rule-based Classifiers	(Xu, Zhang, and Liang 2013), (Alcala-Fdez, Alcala, and Herrera 2011), (Ruiz Reyes et al. 2010), (Temko, Macho, and Nadeu 2008)
	Ensemble Classifier	(Younghyun, Hanseok, and Han 2013), (Dafna, Tarasiuk, and Zigel. 2013), (Li, Wang, and Sung 2008), (Bin, Haizhou, and Rong 2007), (Meyer and Schramm 2006)
	Bayesian Networks	(Giannakopoulos, Pikrakis, and Theodoridis 2007), (Prodanov and Drygajlo 2005), (Daoudi, Fohr, and Antoine 2003), (Zweig 2003)
	Linear Discriminants	(Gergen, Nagathil, and Martin 2014), (Lu and Wang 2012), (Lee et al. 2006)
	Support Vector Machines	(Andreassen, Surlykke, and Hallam 2014), (Muhammad and Melhem 2014), (Costa et al. 2012), (Shuiping, Zhenming, and Shiqiang 2011), (Temko and Nadeu 2009), (Dhanalakshmi, Palanivel, and Ramalingam 2009), (Truong, Lin, and Chen 2007), (Temko and Nadeu 2006), (Acır, Özdamar et al. 2006)
Semi-Supervised Learning Algorithms	Self-training	(Triguero et al. 2014), (Neiberg, Salvi, and Gustafson 2013), (Santos and Canuto 2014), (Yanan et al. 2012)
	Co-training & EM	(Yunyun, Songcan, and Zhi-Hua 2012), (Cui, Jing, and Jen-Tzung 2012), (Yangqiu and Changshui 2008), (Moreno and Agarwal 2003)
	TSVM	(Guz et al. 2010),(Rongyan et al. 2010)

Table 2. Types of datasets for AED.

Dataset Name	Dataset Type	Description
AVICAR (Navarathna et al. 2013)	Speech and Event	AVICAR is a speech dataset recorded using multi-sensory arrays containing four video cameras and eight microphones in a car environment. Speakers of various languages include 50 males and 50 females.
GTZAN (Tsunoo et al. 2011)	Music/Speech	GTZAN is a dataset containing 20 musical genre and 3 speech excerpts with different qualities, each excerpt being 30 seconds long.
Aurora-2 (Truong, Lin, and Chen 2007)	Speech	Aurora-2 is a dataset of recorded hands-free speech to explore the influence on automatic speech recognition performance in noisy situations.
RWCP-DB (Temko and Nadeu 2006)	Event and Environment	RWCP-DB is a dataset with 105 environmental sound events and around 100 anechoic samples of each event in 3 categories: first, collisions such as wood; second, actions like articles dropping; third, characteristics such as small metal articles, paper and instruments.
LVCSR (Meyer and Schramm 2006)	Word Dataset	LVCSR is a rich word dataset when the testing audio is a word stream. This dataset can produce very fast queries with high accuracy, and it is easy to add to, and enhance it to address current issues.
Ballroom (Shen, Shepherd, and Ngu 2006)	Music/Speech	Ballroom is a dataset containing excerpts of many music pieces with real radio quality (low quality)

Figure 1. Block diagram of an audio event detection system.

Figure 2. Feature categorization.

Figure 3. Block diagram of cepstrum computation.

Figure 4. Classification category in audio event detection.

Figure 5. Basic bayesian network structure.

Table 3. Evaluation metrics offered in the AED field.

References Accuracy Metrics Description
(Zubair, Yan, and Wang 2013)		● To calculate the classification accuracy; where is the number of correctly classified samples in one class, is the total number of instances and C is the total number of classes.
(Navarathna et al. 2013)		● To calculate speech word recognition accuracy; where N is the number of words, D is the number of deleted words, I is the number of inserted words and S is the number of substitutions.
(Orio 2010)		● To identify the rate and Mean Reciprocal Rank (MRR); where K is the number of identified items and is the position of item k in the rank list returned by the different approaches.
(Temko and Nadeu 2009)		● To detect all instances; where ACC stands for accuracy and AED for audio event detection, and is a weighting factor that balances Precision and Recall. To score an audio segmentation task; where, and are the segment duration, AEs reference number, AEs output number, and number of references, respectively.

Table 4. Various audio event detection classification approaches.

Methodology	Ref	Detection approach	No. of articles	Technology type	Detection performance	Type of source
Unsupervised Learning Algorithms	(Pomponi and Vinogradov 2013)	Hierarchical and partition clustering	5	Non-speech	Low	Private Dataset
						3
	(Tsunoo et al. 2011)			Non-speech	Moderate	GTZAN & Ballroom Dataset
	(Lefèvre and Vincent 2011)			Non-speech	Moderate	Private Dataset
	(Yang et al. 2013)			Both	High	SCUTMD movie database
	(Park 2009)			Non-speech	High	Private Dataset
						3
	(Chuan 2013)	GMM	5	Speech	High	Monte Carlo Dataset
	(Choi and Chang 2012)			Both	High	NTT Dataset
	(P. Dhanalakshmi et al. 2011b)			Non-speech	High	Private Dataset
				Non-speech		3
	(Cheng, Sun, and Ji 2010)			Non-speech	High	Private Dataset
						3
	(Chung-Hsien and Chia-Hsin 2006)			Speech	Moderate	Private Dataset
				Speech		3
	(Navarathna et al. 2013)	HMM	7	Speech	Low	AVICAR Dataset
	(Niessen, Van Kasteren, and Merentitis 2013)			Both	Low	IEEE Audio Dataset
						16
	(Itoh, Takiguchi, and Ariki 2013)			Non-speech	Moderate	Private Dataset
	(Wang and Zhang 2012)			Non-speech	Moderate	Private Dataset
	(Milone et al. 2012)			Non-speech	High	Private Dataset
	(Ya-Ti et al. 2009)			Non-speech	Moderate	RWCP Dataset
	(Scheme, Hudgins, and Parker 2007)			Speech	High	Private Dataset
						3
	(P. Dhanalakshmi et al. 2011a)	Neural network (SOM, ART)	3	Both	High	Private Dataset
	(Schroeder et al. 2011)			Non-speech	Moderate	Private Dataset
	(Charalampidis, Georgiopoulos, and Kasparis 2000)			Non-speech	Moderate	Private Dataset
Supervised learning Algorithms	(Khunarsal, Lursinsap, and Raicharoen 2013)	KNN	7	Both	High	BBC and Sound Ideas
	(Ravan and Beheshti 2011)				Moderate	CSLU dataset
	(Liu and Zhang 2012)			Non-speech	Moderate	UCI Dataset
	(Huang et al. 2009)			Non-speech	High	Private Dataset
	(Lie, Hong-Jiang, and Hao 2002)			Non-speech	High	Private Dataset
	(Malhotra, Nikolaidis, and Harms 2008)			Both	High	Private Dataset
	(Balochian, Seidabad, and Rad 2013)	Neural Network (RBF – MLP)	7	Both	High	Private Dataset
	(Ganapathy, Rajan, and Hermansky 2011)			Both	Low	NIST Dataset
	(Mitra and Wang 2008)			Non-speech	High	Private Dataset
	(Kotti et al. 2007)			Speech	High	Private Dataset
	(Shen, Shepherd, and Ngu 2006)			Non-speech	High	GTZAN & Ballroom
	(Turnbull and Elkan 2005)			Non-speech	Low	Private Dataset
	(McConaghy et al. 2003)			Non-speech	High	AN/PPS-15
	(Xu, Zhang, and Liang 2013)	Rule-based	5	Non-speech	High	Private Dataset
	(Ruiz Reyes et al. 2010)			Non-speech	High	Private Dataset
	(Temko, Macho, and Nadeu 2008)			Both	High	CLEAR’07 Dataset
	(Hongwei and Mendel 2007)			Non-speech	High	Private Dataset
	(Yaochu 2000)				Low	Private Dataset
	(Dafna, Tarasiuk, and Zigel. 2013)	Ensemble classifier	4	Non-speech	High	Private Dataset
	(Li, Wang, and Sung 2008)			Non-speech	Moderate	Benchmark Dataset
	(Bin, Haizhou, and Rong 2007)			Speech	Moderate	NIST Dataset
	(Meyer and Schramm 2006)			Speech	High	LVCSR Dataset
	(Giannakopoulos, Pikrakis, and Theodoridis 2007)	Bayesian Network	4	Speech	Moderate	Private Dataset
	(Prodanov and Drygajlo 2005)			Speech	Moderate	RoboX Dataset
	(Daoudi, Fohr, and Antoine 2003)			Speech	Moderate	Private Dataset
	(Zweig 2003)			Speech	Moderate	Private Dataset
	(Gergen, Nagathil, and Martin 2014)	LDA	2	Speech	High	TIMIT Dataset
	(Lee et al. 2006)			Non-speech	High	Private Dataset
	(Andreassen, Surlykke, and Hallam 2014)	SVM	9	Non-speech	High	Private Dataset
	(Muhammad and Melhem 2014)			Speech	High	MEEI Dataset
	(Costa et al. 2012)			Non-speech	Moderate	LMD and ISMIR Dataset
	(Shuiping, Zhenming, and Shiqiang 2011)			Non-speech	High	Private Dataset
	(Dhanalakshmi, Palanivel, and Ramalingam 2009)			Both	High	Private Dataset
	(Temko and Nadeu 2009)			Both	Moderate	CLEAR’07
	(Truong, Lin, and Chen 2007)			Speech	High	public audio Dataset
	(Temko and Nadeu 2006)			Both	High	RWCP Dataset
	(Acır, Özdamar et al. 2006)			Non-speech	High	Private Dataset
Semi-Supervised Learning Algorithms		Self-training, EM, TSVM	8
	(Yanan et al. 2012)			Non-speech	High	Private Dataset
	(Santos and Canuto 2014)			Speech	High	Private Dataset
	(Guz et al. 2010)				Moderate	ICSI MRDA Corpus
	(Cui, Jing, and Jen-Tzung 2012)				High	English broadcast news
	(Rongyan et al. 2010)			Speech	High	Friends melodrama Corpus
	(Yangqiu and Changshui 2008)			Non-speech	Moderate	Private Dataset
	(Moreno and Agarwal 2003)			Speech	High	Linguistic Dataset

Table 5. Evaluation of unsupervised learning algorithms.

Ref	Method	Accuracy rate	Type of input
(Choi and Chang 2012)	To detect speech absence and update the GMM likelihood the speech absence probability (Khunarsal, Lursinsap et al.) is employed.	Accuracy of classification: 95.37%	Speech/Non-Speech
(Navarathna et al. 2013)	Mixture of single audio and four visual streams in a five-stream SHMM	Accuracy of recognition: 56%	Speech
(Scheme, Hudgins, and Parker 2007)	HMM is used to classify words at the phoneme level.	Accuracy of Classification over 91% in low noise	Speech
(Cheng, Sun, and Ji 2010)	MFCC and GMM are used across four passerine species.	Accuracy: 89.1–92.5%	Non-speech
(Milone et al. 2012)	HMM is used to classify acoustic signals	Accuracy of recognition: 85%	Non-speech
(P. Dhanalakshmi et al. 2011b)	Auto-associative neural network model (AANN) combined with a GMM-based classifier	Accuracy of classification: 93%	Non-speech
(Dhanalakshmi, Palanivel et al. 2011a)	Mixture of the auto-associative neural network model (AANN) with GMM	Accuracy of classification: 93.0%	Speech/Non-speech
(Yang et al. 2013)	Spectral clustering and k-means to cluster audio events	Accuracy of detection: %88.63	Speech/Non-speech
(Park 2009)	FCM-DK relies on the fuzzy c-means algorithm that uses a kernel method for data transformation.	Accuracy of classification: 89.12%,	Non-Speech
(Chung-Hsien and Chia-Hsin 2006)	A minimum description length (MDL)-based GMM statistically designates the audio features.	False alarm rate: 0.14 Accuracy of classification: 88%	Speech
(Chuan 2013)	Discrete wavelet transform is used to extract audio features at different scales and time from audio recordings.	Accuracy of classification: 87.32%	Speech/Non-speech
(Niessen, Van Kasteren, and Merentitis 2013)	A hierarchical HMM for sound event detection	F-measure recognition: 45.5%	Speech/Non-speech
(Itoh, Takiguchi, and Ariki 2013)	HMM is used.	Precision: 66.67%, Recall: 100%	Non-Speech
(Wang and Zhang 2012)	HMM mixture to create a framework for ice hockey videos.	Accuracy of HMM 73.01%, Gaussian mixture HMM 59.15%	Non-speech
(Ya-Ti et al. 2009)	A fusion method mixed with hierarchical HMM is employed to cover large differences existing in healthcare audio events.	Accuracy of cassification: 70%	Non-speech
(Schroeder et al. 2011)	Technical method employed in the context of ambient assisted living (AAL).	TPR over 79%, FPR smaller than 4%	Non-speech
(Tsunoo et al. 2011)	K-means algorithm based on one-pass programming.	Accuracy on GTZAN and ballroom dataset 72.4% and 52.5% respectively	Non-speech
(Lefèvre and Vincent 2011)	The proposed method depends on classical HMM, cepstral or spectral analysis and amplitude.	For whistle, crowd and speaker, recall achieved is 95%, 75%, and 95% and precision is 86%, 86%, and 90%	Non-speech

Figure 6. Year versus distribution of articles on different classifier types.

Figure 7. Comparison of unsupervised learning algorithms in terms of accuracy rate.

Table 6. Evaluation of supervised learning algorithms.

Ref	Method	Accuracy Rate	Type of input
(Ganapathy, Rajan, and Hermansky 2011)	Speech activity detection (Lin, Li et al.) techniques with an MLP.	Equal error rate: 9%	Speech/Non-speech
(Huang et al. 2009)	K-NN and SVM are employed with three different features to detect frog type	Accuracy: 89.67%	Non-speech
(Xu, Zhang, and Liang 2013)	Wavelet Packet Transform method and Fuzzy SVM	Accuracy of identification: 99%,	Non-speech
(Khunarsal, Lursinsap, and Raicharoen 2013)	Spectrogram pattern matching along with neural network and KNN classifiers	Accuracy: 90.57%.	Speech/Non-speech
(Ravan and Beheshti 2011)	The power spectrum density (PSD) of each speech signal frame is estimated using KNN classification	Accuracy: 90%	Speech
(Liu and Zhang 2012)	New anomaly removal and learning algorithm under the KNN framework	Accuracy: 99.3% when k = 3	Noisy data
(Lie, Hong-Jiang, and Hao 2002)	KNN and linear spectral pairs-vector quantization (LSP-VQ) to discriminate speech from non-speech	Accuracy: over 96%.	Speech/Non-speech
(Balochian, Seidabad, and Rad 2013)	Optimized MLP classifiers to execute some features based on the wavelet transform	Accuracy: 96.49%	Speech/Non-speech
(Kotti et al. 2007)	The cross-correlation function and magnitude of the corresponding cross-power spectral density are fed as input to the neural network for recognition.	Accuracy: 88.1%	Speech
(Shen, Shepherd, and Ngu 2006)	Combination of multiple musical characteristics with a hybrid architecture based on principal component analysis (PCA) and MLP	Accuracy: 91.9%	Non-speech
(Turnbull and Elkan 2005)	Achieved with RBF networks by using a combination of unsupervised and supervised initialization methods	Accuracy: 71.5%	Non-speech
(McConaghy et al. 2003)	RBF algorithms are employed to classify real-life audio radar signals	Accuracy: 86.4%.	Non-speech
(Ruiz Reyes et al. 2010)	Genetic fuzzy system	Accuracy of classification: 97%	Speech/Non-speech
(Temko, Macho, and Nadeu 2008)	Fusion of different information sources with fuzzy integral (FI), fuzzy measure (FM), and SVM	Accuracy 83.9% and precision 81.2%	Speech/Non-speech
(Dafna, Tarasiuk, and Zigel. 2013)	AdaBoost-based algorithms for detection of snore/non-snore	Accuracy of detection: 98.2%	Speech/Non-speech
(Li, Wang, and Sung 2008)	AdaBoostSVM used for sequence of trained RBFSVM classifiers	Accuracy: 85%	Non-speech
(Bin, Haizhou, and Rong 2007)	Ensemble of binary classifiers.	Equal error rate: 1.38% and 3.20%	Speech
(Meyer and Schramm 2006)	AdaBoost.M2 is used for HMM-based speech recognition	Equal error rate: 0.8%	Speech
(Giannakopoulos, Pikrakis, and Theodoridis 2007)	BN in combination with One Versus All classification architecture.	Accuracy: 73.2%, False alarm rate: 11%.	Speech/Non-speech
(Prodanov and Drygajlo 2005)	Bayesian network framework	Accuracy: 77.9%	Speech
(Daoudi, Fohr, and Antoine 2003)	A speech model is built in the time–frequency domain using the formalism of dynamic BNs	Accuracy: 87.89%	Speech
(Zweig 2003)	BNs are used to model a wide variety of phenomena that occur in speech recognition.	Error rate 3.1	Speech
(Gergen, Nagathil, and Martin 2014)	Analyze the influence of reverberation and competing acoustical sources on the classification of audio signals captured by ad hoc distributed microphones	Accuracy: 90%	Speech
(Lee et al. 2006)	UMFCCs and LDA	Accuracy: 96.8% and 98.1%	Non-speech
(Andreassen, Surlykke, and Hallam 2014)	SVM is used based on a combination of temporal and spectral analyses to classify events	Accuracy 96% for dry nights and 70% when raining	Non-speech
(Muhammad and Melhem 2014)	MPEG-7 features are used for indexing, including both video and audio	Accuracy: 99.994%	Non-speech
(Costa et al. 2012)	The audio signal is converted to a spectrogram and features are extracted from time-frequency	Accuracy: 82.33%	Non-speech
(Shuiping, Zhenming, and Shiqiang 2011)	MFFC, ZCR, etc., features are extracted and an audio classification based on SVM was designed	Accuracy: 90%.	Non-speech
(Temko and Nadeu 2009)	SVM-based two-step system outperforms the baseline system for an artificially-generated database	Accuracy: 90.30%	Speech/Non-speech
(Dhanalakshmi, Palanivel, and Ramalingam 2009)	Support vector machines are applied with neural networks (RBFNN)	Accuracy of classification: 92%	Speech/Non-speech
(Truong, Lin, and Chen 2007)	Wavelets and SVMs are employed to segment specific speakers	Accuracy of 94.12% and 85.93% for4-speaker and8-speaker	Speech
(Temko and Nadeu 2006)	Several classifiers based on SVM using confusion matrix-based clustering schemes to deal with multi-class problems	Accuracy 88.29%, average error reduction 31.5%	Non-speech
(Acır, Özdamar et al. 2006)	An SVM classifier is employed with discrete cosine transform (DCT) coefficients and discrete wavelet transform (DWT)	Sensitivity 95.3%, specificity 84.6% and accuracy 93.8%.	Non-speech

Table 7. Evaluation of semi-supervised learning algorithms.

Ref	Method	Accuracy Rate	Type of input
(Moreno and Agarwal 2003)	EM-based algorithms for semi-supervised learning with a Gaussian mixture model	Error reduction: 5.04%	Speech
(Cui, Jing, andJen-Tzung 2012)	Cross-view transfer learning for LVCSR through a committee machine	Accuracy: 82.95%	Speech
(Yangqiu and Changshui 2008)	EM algorithm to adaptively learn the fusion scores	Accuracy: 75%	Non-speech
(Rongyan et al. 2010)	TSVM algorithm for automatic AE annotation	Accuracy: 89.6%	Speech/Non-speech
(Guz et al. 2010)	Self-training and co-training classification using lexical and prosodic information	Accuracy: 74.2%	Speech
(Li, Zhang, and Ma 2012)	The CBSL algorithm first selects the sentence level training corpus and then introduces the confirmation criterion to select the word level corpus	Accuracy: 98.83%	Speech
(Yanan et al. 2012)	A semi-supervised gait recognition algorithm based on self-training	Accuracy: 92.4%	Non-speech

Figure 8. Comparison of supervised learning algorithms in terms of accuracy rate.

Figure 9. Comparison of semi-supervised learning algorithms in terms of accuracy rate.

Figure 10. Comparison of supervised, unsupervised and semi-supervised learning algorithms in terms of accuracy rate.

Pomponi, E., and A. Vinogradov. 2013. A real-time approach to acoustic emission clustering. Mechanical Systems and Signal Processing 40 (2):791–804.

 Web of Science ®Google Scholar

Tsunoo, E., G. Tzanetakis, N. Ono, and S. Sagayama. 2011. Beyond timbral statistics: Improving music classification using percussive patterns and bass lines. IEEE Transactions on Audio, Speech, and Language Processing 19 (4):1003–14.

 Google Scholar

Park, D.-C. 2009. Classification of audio signals using Fuzzy c-means with divergence-based Kernel. Pattern Recognition Letters 30 (9):794–98.

 Web of Science ®Google Scholar

Lefèvre, S., and N. Vincent. 2011. A two level strategy for audio segmentation. Digital Signal Processing 21 (2):270–77.

 Web of Science ®Google Scholar

Yang, J.-C., L.-A. Liu, Q.-W. Qin, and M. Zhang. 2013. Audio event change detection and clustering in movies. Journal of Multimedia 8 (2):113–20.

 Google Scholar

Chuan, C.-H. 2013. Audio classification and retrieval using wavelets and gaussian mixture models. International Journal Multimed Data Engineering Managed 4 (1):1–20.

 Google Scholar

Choi, J.-H., and J.-H. Chang. 2012. On using acoustic environment classification for statistical model-based speech enhancement. Speech Communication 54 (3):477–90.

 Web of Science ®Google Scholar

Dhanalakshmi, P., S. Palanivel, and V. Ramalingam. 2011b. Pattern classification models for classifying and indexing audio signals. Engineering Applications of Artificial Intelligence 24 (2):350–57.

 Web of Science ®Google Scholar

Cheng, J., Y. Sun, and L. Ji. 2010. A call-independent and automatic acoustic system for the individual recognition of animals: A novel model using four passerines. Pattern Recognition 43 (11):3846–52.

 Web of Science ®Google Scholar

Chung-Hsien, W., and H. Chia-Hsin. 2006. Multiple change-point audio segmentation and classification using an MDL-based Gaussian model. IEEE Transactions on Audio, Speech, and Language Processing 14 (2):647–57.

 Google Scholar

Navarathna, R., D. Dean, S. Sridharan, and P. Lucey. 2013. Multiple cameras for audio-visual speech recognition in an automotive environment. Computer Speech & Language 27 (4):911–27.

 Web of Science ®Google Scholar

Niessen, M. E., T. L. M. Van Kasteren, and A. Merentitis. 2013. Hierarchical modeling using automated sub-clustering for sound event recognition. IEEE Workshop on Applications of Signal Processing to Audio and Acoustics (WASPAA).

 Google Scholar

Itoh, H., T. Takiguchi, and Y. Ariki. 2013. Event detection and recognition using HMM with whistle sounds. Signal-Image Technology & Internet-Based Systems (SITIS), 2013 International Conference on.

 Google Scholar

Wang, X., and X.-P. Zhang. 2012. Ice hockey shooting event modeling with mixture hidden Markov model. Multimedia Tools and Applications 57 (1):131–44.

 Web of Science ®Google Scholar

Milone, D. H., J. R. Galli, C. A. Cangiano, H. L. Rufiner, and E. A. Laca. 2012. Automatic recognition of ingestive sounds of cattle based on hidden Markov models. Computers and Electronics in Agriculture 87:51–55.

 Web of Science ®Google Scholar

Ya-Ti, P., L. Ching-Yung, S. Ming-Ting, and T. Kun-Cheng. 2009. Healthcare audio event classification using hidden markov models and hierarchical hidden markov models. IEEE International Conference on Multimedia and Expo 2009. ICME 2009.

 Google Scholar

Scheme, E. J., B. Hudgins, and P. A. Parker. 2007. Myoelectric signal classification for phoneme-based speech recognition. IEEE Transactions on Biomedical Engineering 54 (4):694–99.

 PubMed Web of Science ®Google Scholar

Dhanalakshmi, P., S. Palanivel, and V. Ramalingam. 2011a. Classification of audio signals using AANN and GMM. Applied Soft Computing 11 (1):716–23.

 Web of Science ®Google Scholar

Schroeder, J., S. Wabnik, P. J. Hengel, and S. Goetze. 2011. Detection and classification of acoustic events for in-home care. In Ambient assisted living, Eds. R. Wichert, and B. Eberhardt, Springer Berlin Heidelberg, 181–195.

 Google Scholar

Charalampidis, D., M. Georgiopoulos, and T. Kasparis. 2000. Classification of noisy signal using fuzzy ARTMAP neural networks. International Joint Conference on Neural Networks, IJCNN2000, Proceedings of the IEEE-INNS-ENNS.

 Google Scholar

Khunarsal, P., C. Lursinsap, and T. Raicharoen. 2013. Very short time environmental sound classification based on spectrogram pattern matching. Information Sciences 243:57–74.

 Web of Science ®Google Scholar

Ravan, M., and S. Beheshti. 2011. Speech recognition from adaptive windowing PSD estimation. 24th Canadian Conference on Electrical and Computer Engineering (CCECE), 524–27.

 Google Scholar

Liu, H., and S. Zhang. 2012. Noisy data elimination using mutual k-nearest neighbor for classification mining. Journal of Systems and Software 85 (5):1067–74.

 Web of Science ®Google Scholar

Lie, L., Z. Hong-Jiang, and J. Hao. 2002. Content analysis for audio classification and segmentation. IEEE Transactions on Speech and Audio Processing 10 (7):504–16.

 Google Scholar

Huang, C.-J., Y.-J. Yang, D.-X. Yang, and Y.-J. Chen. 2009. Frog classification using machine learning techniques. Expert Systems with Applications 36 (2, Part 2):3737–43.

 Web of Science ®Google Scholar

Malhotra, B., I. Nikolaidis, and J. Harms. 2008. Distributed classification of acoustic targets in wireless audio-sensor networks.”. Computation Network 52 (13):2582–93.

 Web of Science ®Google Scholar

Balochian, S., E. A. Seidabad, and S. Z. Rad. 2013. Neural network optimization by genetic algorithms for the audio classification to speech and music. International Journal of Signal Processing, Image Processing & Pattern Recognition 6 (3).

 Google Scholar

Ganapathy, S., P. Rajan, and H. Hermansky. 2011. Multi-layer perceptron based speech activity detection for speaker verification. IEEE Workshop on Applications of Signal Processing to Audio and Acoustics (WASPAA).

 Google Scholar

Mitra, V., and C.-J. Wang. 2008. Content based audio classification: A neural network approach. Soft Computing 12 (7):639–46.

 Web of Science ®Google Scholar

Kotti, M., E. Benetos, C. Kotropoulos, and I. Pitas. 2007. A neural network approach to audio-assisted movie dialogue detection. Neurocomputing 71 (1–3):157–66.

 Web of Science ®Google Scholar

Shen, J., J. Shepherd, and A. H. H. Ngu. 2006. Towards effective content-based music retrieval with multiple acoustic feature combination. IEEE Transactions on Multimedia 8 (6):1179–89.

 Web of Science ®Google Scholar

Turnbull, D., and C. Elkan. 2005. Fast recognition of musical genres using RBF networks. IEEE Transactions on Knowledge and Data Engineering 17 (4):580–84.

 Web of Science ®Google Scholar

Khairnar, D. G., S. N. Merchant, and U. B. Desai. 2005. An optimum RBF network for signal detection in non-gaussian noise. In Pattern recognition and machine intelligence, Eds. S. Pal, S. Bandyopadhyay, and S. Biswas, Springer Berlin Heidelberg, 3776:306–309.

 Google Scholar

McConaghy, T., H. Leung, E. Bosse, and V. Varadan. 2003. Classification of audio radar signals using radial basis function neural networks. IEEE Transactions on Instrumentation and Measurement 52 (6):1771–79.

 Web of Science ®Google Scholar

Xu, Q., L. Zhang, and W. Liang. 2013. Acoustic detection technology for gas pipeline leakage. Process Safety and Environmental Protection 91 (4):253–61.

 Web of Science ®Google Scholar

Alcala-Fdez, J., R. Alcala, and F. Herrera. 2011. A fuzzy association rule-based classification model for high-dimensional problems with genetic rule selection and lateral tuning. IEEE Transactions on Fuzzy Systems 19 (5):857–72.

 Web of Science ®Google Scholar

Ruiz Reyes, N., P. Vera Candeas, S. García Galán, and J. E. Muñoz. 2010. Two-stage cascaded classification approach based on genetic fuzzy learning for speech/music discrimination. Engineering Applications of Artificial Intelligence 23 (2):151–59.

 Web of Science ®Google Scholar

Temko, A., D. Macho, and C. Nadeu. 2008. Fuzzy integral based information fusion for classification of highly confusable non-speech sounds. Pattern Recognition 41 (5):1814–23.

 Web of Science ®Google Scholar

Younghyun, L., K. Hanseok, and D. K. Han. 2013. Acoustic signal based abnormal event detection system with multiclass adaboost. IEEE International Conference on Consumer Electronics (ICCE), 2013.

 Google Scholar

Dafna, E., A. Tarasiuk, and Y. Zigel. 2013. Automatic detection of whole night snoring events using non-contact microphone. PLoS One 8 (12). doi:10.1371/journal.pone.0084139

 PubMed Web of Science ®Google Scholar

Li, X., L. Wang, and E. Sung. 2008. AdaBoost with SVM-based component classifiers. Engineering Applications of Artificial Intelligence 21 (5):785–95.

 Web of Science ®Google Scholar

Bin, M., L. Haizhou, and T. Rong. 2007. Spoken language recognition using ensemble classifiers. IEEE Transactions on Audio, Speech, and Language Processing 15 (7):2053–62.

 Google Scholar

Meyer, C., and H. Schramm. 2006. Boosting HMM acoustic models in large vocabulary speech recognition. Speech Communication 48 (5):532–48.

 Web of Science ®Google Scholar

Giannakopoulos, T., A. Pikrakis, and S. Theodoridis. 2007. A multi-class audio classification method with respect to violent content in movies using bayesian networks. Multimedia Signal Processing, 2007. MMSP 2007. IEEE 9th Workshop on.

 Google Scholar

Prodanov, P., and A. Drygajlo. 2005. Bayesian networks based multi-modality fusion for error handling in human–Robot dialogues under noisy conditions. Speech Communication 45 (3):231–48.

 Web of Science ®Google Scholar

Daoudi, K., D. Fohr, and C. Antoine. 2003. Dynamic Bayesian networks for multi-band automatic speech recognition. Computer Speech & Language 17 (2–3):263–85.

 Web of Science ®Google Scholar

Zweig, G. 2003. Bayesian network structures and inference techniques for automatic speech recognition. Computer Speech & Language 17 (2–3):173–93.

 Web of Science ®Google Scholar

Gergen, S., A. Nagathil, and R. Martin. 2014. Classification of reverberant audio signals using clustered ad hoc distributed microphones. Signal Processing 107:21–32.

 Web of Science ®Google Scholar

Lu, G.-F., and Y. Wang. 2012. Feature extraction using a fast null space based linear discriminant analysis algorithm. Information Sciences 193:72–80.

 Web of Science ®Google Scholar

Lee, C.-H., C.-H. Chou, -C.-C. Han, and R.-Z. Huang. 2006. Automatic recognition of animal vocalizations using averaged MFCC and linear discriminant analysis. Pattern Recognition Letters 27 (2):93–101.

 Web of Science ®Google Scholar

Andreassen, T., A. Surlykke, and J. Hallam. 2014. Semi-automatic long-term acoustic surveying: A case study with bats. Ecological Informatics 21:13–24.

 Web of Science ®Google Scholar

Muhammad, G., and M. Melhem. 2014. Pathological voice detection and binary classification using MPEG-7 audio features. Biomedical Signal Processing and Control 11:1–9.

 Web of Science ®Google Scholar

Costa, Y. M. G., L. S. Oliveira, A. L. Koerich, F. Gouyon, and J. G. Martins. 2012. Music genre classification using LBP textural features. Signal Processing 92 (11):2723–37.

 Web of Science ®Google Scholar

Shuiping, W., T. Zhenming, and L. Shiqiang. 2011. Design and implementation of an audio classification system based on SVM. Procedia Engineering 15:4031–35.

 Google Scholar

Temko, A., and C. Nadeu. 2009. Acoustic event detection in meeting-room environments. Pattern Recognition Letters 30 (14):1281–88.

 Web of Science ®Google Scholar

Dhanalakshmi, P., S. Palanivel, and V. Ramalingam. 2009. Classification of audio signals using SVM and RBFNN. Expert Systems with Applications 36 (3):6069–75.

 Web of Science ®Google Scholar

Truong, T. K., -C.-C. Lin, and S.-H. Chen. 2007. Segmentation of specific speech signals from multi-dialog environment using SVM and wavelet. Pattern Recognition Letters 28 (11):1307–13.

 Web of Science ®Google Scholar

Temko, A., and C. Nadeu. 2006. Classification of acoustic events using SVM-based clustering schemes. Pattern Recognition 39 (4):682–94.

 Web of Science ®Google Scholar

Acır, N., Ö. Özdamar, and C. Güzeliş. 2006. Automatic classification of auditory brainstem responses using SVM-based feature selection algorithm for threshold detection. Engineering Applications of Artificial Intelligence 19 (2):209–18.

 Web of Science ®Google Scholar

Triguero, I., J. A. Sáez, J. Luengo, S. García, and F. Herrera. 2014. On the characterization of noise filters for self-training semi-supervised in nearest neighbor classification. Neurocomputing 132:30–41.

 Web of Science ®Google Scholar

Neiberg, D., G. Salvi, and J. Gustafson. 2013. Semi-supervised methods for exploring the acoustics of simple productive feedback. Speech Communication 55 (3):451–69.

 Web of Science ®Google Scholar

Santos, A., and A. Canuto. 2014. Applying semi-supervised learning in hierarchical multi-label classification. Expert Systems with Applications 41(14): 6075–6085.

 PubMed Web of Science ®Google Scholar

Yanan, L., Y. Yilong, L. Lili, P. Shaohua, and Y. Qiuhong. 2012. Semi-supervised gait recognition based on self-training. IEEE Ninth International Conference on Advanced Video and Signal-Based Surveillance (AVSS), 2012.

 Google Scholar

Yunyun, W., C. Songcan, and Z. Zhi-Hua. 2012. New semi-supervised classification method based on modified cluster assumption. IEEE Transactions on Neural Networks and Learning Systems 23 (5):689–702.

 PubMed Web of Science ®Google Scholar

Cui, X., H. Jing, and C. Jen-Tzung. 2012. Multi-view and multi-objective semi-supervised learning for HMM-based automatic speech recognition. IEEE Transactions on Audio, Speech, and Language Processing 20 (7):1923,1935.

 Google Scholar

Yangqiu, S., and Z. Changshui. 2008. Content-based information fusion for semi-supervised music genre classification. IEEE Transactions on MultiMedia 10 (1):145–52.

 Web of Science ®Google Scholar

Moreno, P. J., and S. Agarwal. 2003. An experimental study of EM-based algorithms for semi-supervised learning in audio classification. ICML 2003 Workshop on the Continuum from Labeled to Unlabeled Data in Machine Learning and Data Mining.

 Google Scholar

Guz, U., S. Cuendet, D. Hakkani-Tür, and G. Tur. 2010. Multi-view semi-supervised learning for dialog act segmentation of speech. IEEE Transactions on Audio, Speech, and Language Processing 18 (2):320,329.

 Google Scholar

Rongyan, W., L. Gang, G. Jun, and M. Zhenxin. 2010. Semi-supervised learning for automatic audio events annotation using TSVM. International Conference on Computer Application and System Modeling (ICCASM).

 Google Scholar

Zubair, S., F. Yan, and W. Wang. 2013. Dictionary learning based sparse coefficients for audio classification with max and average pooling. Digital Signal Processing 23 (3):960–70.

 Web of Science ®Google Scholar

Orio, N. 2010. Automatic identification of audio recordings based on statistical modeling. Signal Processing 90 (4):1064–76.

 Web of Science ®Google Scholar

Hongwei, W., and J. M. Mendel. 2007. Classification of battlefield ground vehicles using acoustic features and fuzzy logic rule-based classifiers. IEEE Transactions on Fuzzy Systems 15 (1):56–72.

 Web of Science ®Google Scholar

Yaochu, J. 2000. Fuzzy modeling of high-dimensional systems: Complexity reduction and interpretability improvement. IEEE Transactions on Fuzzy Systems 8 (2):212–21.

 Web of Science ®Google Scholar

Li, H., T. Zhang, and L. Ma. 2012. Confirmation based self-learning algorithm in LVCSR’s semi-supervised incremental learning. Procedia Engineering 29:754–59.

 Google Scholar