Acoustic novelty detection with adversarial autoencoders

Principi, Emanuele; Vesperini, Fabio; Squartini, Stefano; Piazza, Francesco

doi:10.1109/ijcnn.2017.7966273

Cited by 40 publications

(43 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We used data from 1,555 feeding executions collected from 24 able-bodied participants where we newly collected 1,203 non-anomalous feeding executions for this work. 16 participants were male and 8 were female, and the age range was [19][20][21][22][23][24][25][26][27][28][29][30][31][32][33][34][35]. We conducted the studies with approval from the Georgia Tech Institutional Review Board (IRB).…”

Section: B Data Collectionmentioning

confidence: 99%

A Multimodal Anomaly Detector for Robot-Assisted Feeding Using an LSTM-Based Variational Autoencoder

Park

Hoshi

Kemp

2018

IEEE Robot. Autom. Lett.

606

283

View full text Add to dashboard Cite

The detection of anomalous executions is valuable for reducing potential hazards in assistive manipulation. Multimodal sensory signals can be helpful for detecting a wide range of anomalies. However, the fusion of high-dimensional and heterogeneous modalities is a challenging problem. We introduce a long short-term memory based variational autoencoder (LSTM-VAE) that fuses signals and reconstructs their expected distribution. We also introduce an LSTM-VAE-based detector using a reconstruction-based anomaly score and a state-based threshold. For evaluations with 1,555 robot-assisted feeding executions including 12 representative types of anomalies, our detector had a higher area under the receiver operating characteristic curve (AUC) of 0.8710 than 5 other baseline detectors from the literature. We also show the multimodal fusion through the LSTM-VAE is effective by comparing our detector with 17 raw sensory signals versus 4 hand-engineered features.

show abstract

Section: B Data Collectionmentioning

confidence: 99%

A Multimodal Anomaly Detector for Robot-Assisted Feeding Using an LSTM-Based Variational Autoencoder

Park

Hoshi

Kemp

2018

IEEE Robot. Autom. Lett.

606

283

View full text Add to dashboard Cite

show abstract

“…The dataset was recorded by a binaural microphone at a sample rate of 16kHz. We converted each audio to 1 channel and then split it into sequences of 160-dimensional frames, each frame corresponds to 0.01s, as in [12] and [13]. [12] and [13] evaluated the detection at each frame instead of at the whole sequence, so we also applied the thresholding step to each log p(x t |x <t ), instead of log p(x 1:T ).…”

Section: Methodsmentioning

confidence: 99%

“…Both [12] and [13] used RNNs (LSTMs in particular) as an AutoEncoder (AE) which can reconstruct the original signal from a compressed representation (Compression AutoEncoders -CAEs) or from a corrupted version of it (Denoising AutoEncoders -DAEs). However, as discussed in [16], [17] and [20], the fact that the hidden states of RNNs are deterministic reduces their capacity to capture all data variabilities, especially for data that contain high levels of randomness.…”

Section: Related Workmentioning

confidence: 99%

Recurrent Neural Networks with Stochastic Layers for Acoustic Novelty Detection

Nguyen

Kirsebom

Frazão

et al. 2019

ICASSP 2019 - 2019 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

View full text Add to dashboard Cite

In this paper, we adapt Recurrent Neural Networks with Stochastic Layers, which are the state-of-the-art for generating text, music and speech, to the problem of acoustic novelty detection. By integrating uncertainty into the hidden states, this type of network is able to learn the distribution of complex sequences. Because the learned distribution can be calculated explicitly in terms of probability, we can evaluate how likely an observation is then detect low-probability events as novel. The model is robust, highly unsupervised, end-toend and requires minimum preprocessing, feature engineering or hyperparameter tuning. An experiment on a benchmark dataset shows that our model outperforms the state-of-the-art acoustic novelty detectors.Index Termsacoustic modeling, novelty detection, variational recurrent neural network, stochastic recurrent neural network.

show abstract

“…Due to the inherent potential for capturing data distributions, there is a growing body of literature that recognizes the importance of AAE. In [41], Principi et al proposed an acoustic novelty detector based on AAE, and the results showed that the proposed approach provides a relative performance improvement equal to 0.26% compared to the standard autoencoder. A conditional difference adversarial autoencoder (CDAAE) [42] was proposed for facial expression synthesis to handle the problem of disambiguating changes.…”

Section: Related Workmentioning

confidence: 99%

Background Learning Based on Target Suppression Constraint for Hyperspectral Target Detection

Xie

Zhang

et al. 2020

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

View full text Add to dashboard Cite

Hyperspectral target detection is critical in both military and civilian applications. However, it is a challenging task due to the complexity of background and the limited samples of target in hyperspectral images (HSIs). In this paper, we propose a novel background learning model, called background learning based on target suppression constraint (BLTSC) to characterize high-dimensional spectral vectors. Considering insufficient target samples, the model is trained only on the background spectral samples to accurately learn the background distribution. Then the discrepancy between the reconstructed and original HSIs are examined to spot the targets. To obtain a background training dataset, coarse detection is carried out. However, it is quite difficult to retrieve pure background data. Thus a target suppression constraint is imposed to reduce the impact of suspected target samples on background reconstruction. Experiments on six real HSIs demonstrate that the proposed framework significantly outperforms the current state-of-the-art detection methods and yields higher detection accuracy and lower false alarm rate. Index Terms-Hyperspectral image (HSI), target detection, background learning, target suppression constraint.

show abstract

Acoustic novelty detection with adversarial autoencoders

Cited by 40 publications

References 25 publications

A Multimodal Anomaly Detector for Robot-Assisted Feeding Using an LSTM-Based Variational Autoencoder

A Multimodal Anomaly Detector for Robot-Assisted Feeding Using an LSTM-Based Variational Autoencoder

Recurrent Neural Networks with Stochastic Layers for Acoustic Novelty Detection

Background Learning Based on Target Suppression Constraint for Hyperspectral Target Detection

Contact Info

Product

Resources

About