Comparing the Influence of Depth and Width of Deep Neural Network Based on Fixed Number of Parameters for Audio Event Detection

2018 ◽  
Author(s):  
Jun Wang ◽  
Shengchen Li
Keyword(s):  
Neural Network ◽  
Event Detection ◽  
Deep Neural Network ◽  
Fixed Number ◽  
Audio Event

Using multi-stream hierarchical deep neural network to extract deep audio feature for acoustic event detection

2017 ◽  
Vol 77 (1) ◽  
pp. 897-916 ◽  
Author(s):  
Yanxiong Li ◽  
Xue Zhang ◽  
Hai Jin ◽  
Xianku Li ◽  
Qin Wang ◽  
...  
Keyword(s):  
Neural Network ◽  
Event Detection ◽  
Deep Neural Network ◽  
Acoustic Event Detection ◽  
Acoustic Event ◽  
Audio Feature

Normal and Abnormal Non-Speech Audio Event Detection Using MFCC and PR-Based Feature Sets

2012 ◽  
Vol 601 ◽  
pp. 200-208 ◽  
Author(s):  
Buket D. Barkana ◽  
Burak Uzkent ◽  
Inci Saricicek
Keyword(s):  
Neural Network ◽  
Event Detection ◽  
Context Awareness ◽  
Rain Event ◽  
Pitch Range ◽  
Feature Sets ◽  
Audio Surveillance ◽  
Active Subject ◽  
Audio Event ◽  
Artificial Neural Network Ann

Non-speech audio event detection and classification has become a very active subject of research, since it can be implemented in many important areas: audio surveillance and context awareness systems. In this study, non-speech normal and abnormal audio events were detected by Mel-frequency cepstrum coefficients (MFCC) and Pitch range (PR) based features using artificial neural network (ANN) classifiers. We have 4 abnormal events (glass breaking, dog barking, scream, gunshot) and 2 normal events (engine noise and rain). Event detection, using ANN classifiers, resulted in an accuracy of up to 92%, with recognition rates overall in the range of 78%-87.5%.

Real-Time Multi-Event Anomaly Detection using Elliptic Envelope and A Deep Neural Network for Enhanced MPD Robustness

2021 ◽  
Author(s):  
Qifan Gu ◽  
Amirhossein Fallah ◽  
Pradeepkumar Ashok ◽  
Dongmei Chen ◽  
Eric Van Oort
Keyword(s):  
Neural Network ◽  
Event Detection ◽  
Deep Neural Network ◽  
Change Point Detection ◽  
False Alarms ◽  
Two Phase ◽  
Signal Noise ◽  
Lost Circulation ◽  
Gas Kick ◽  
Few Data

Abstract In managed pressure drilling (MPD), robust and fast event detection is critical for timely event identification and diagnosis, as well as executing well control actions as quickly as possible. In current event detection systems (EDSs), signal noise and uncertainties often cause missed and false alarms, and automated diagnosis of the event type is usually restricted to certain event types. A new EDS method is proposed in this paper to overcome these shortcomings. The new approach uses a multivariate online change point detection (OCPD) method based on elliptic envelope for event detection. The method is robust against signal noise and uncertainties, and is able to detect abnormal features within a minute or less, using only a few data points. A deep neural network (DNN) is utilized for estimating the occurrence probability of various drilling events, currently encompassing (but not limited to) six event types: liquid kick, gas kick, lost circulation, plugged choke, plugged bit, and drillstring washout. The OCPD and the DNN are integrated together and demonstrate better performance with respect to robustness and accuracy. The training and testing of the OCPD and the DNN were conducted on a large dataset representing various drilling events, which was generated using a field-validated two-phase hydraulics software. Compared to current EDS methods, the new system shows the following advantages: (1) lower missed alarm rate; (2) lower false alarm rate; (3) earlier alarming; and (4) significantly improved classification capability that also allows for further extension to even more drilling events.

Event Detection and Control of Blood Glucose Levels Using Deep Neural Network

2021 ◽  
pp. 235-245
Author(s):  
Divya Govindaraju ◽  
R. R. Rashmika Shree ◽  
S. Priyanka ◽  
S. Porkodi ◽  
Sutha Subbian
Keyword(s):  
Neural Network ◽  
Blood Glucose ◽  
Event Detection ◽  
Deep Neural Network ◽  
Glucose Levels ◽  
Blood Glucose Levels ◽  
And Control

scholarly journals Joint Optimization of Deep Neural Network-Based Dereverberation and Beamforming for Sound Event Detection in Multi-Channel Environments

Sensors ◽  
2020 ◽  
Vol 20 (7) ◽  
pp. 1883 ◽  
Author(s):  
Kyoungjin Noh ◽  
Joon-Hyuk Chang
Keyword(s):  
Neural Network ◽  
Loss Function ◽  
Event Detection ◽  
Deep Neural Network ◽  
Joint Optimization ◽  
Audio Signals ◽  
Sound Event ◽  
Minimum Variance Distortionless Response ◽  
Reverberant Environments ◽  
Sound Event Detection

In this paper, we propose joint optimization of deep neural network (DNN)-supported dereverberation and beamforming for the convolutional recurrent neural network (CRNN)-based sound event detection (SED) in multi-channel environments. First, the short-time Fourier transform (STFT) coefficients are calculated from multi-channel audio signals under the noisy and reverberant environments, which are then enhanced by the DNN-supported weighted prediction error (WPE) dereverberation with the estimated masks. Next, the STFT coefficients of the dereverberated multi-channel audio signals are conveyed to the DNN-supported minimum variance distortionless response (MVDR) beamformer in which DNN-supported MVDR beamforming is carried out with the source and noise masks estimated by the DNN. As a result, the single-channel enhanced STFT coefficients are shown at the output and tossed to the CRNN-based SED system, and then, the three modules are jointly trained by the single loss function designed for SED. Furthermore, to ease the difficulty of training a deep learning model for SED caused by the imbalance in the amount of data for each class, the focal loss is used as a loss function. Experimental results show that joint training of DNN-supported dereverberation and beamforming with the SED model under the supervision of focal loss significantly improves the performance under the noisy and reverberant environments.

scholarly journals Deep Convolutional Neural Network with Structured Prediction for Weakly Supervised Audio Event Detection

2019 ◽  
Vol 9 (11) ◽  
pp. 2302 ◽  
Author(s):  
Inkyu Choi ◽  
Soo Hyun Bae ◽  
Nam Soo Kim
Keyword(s):  
Neural Network ◽  
Convolutional Neural Network ◽  
Event Detection ◽  
Large Scale ◽  
Prediction Method ◽  
Deep Convolutional Neural Network ◽  
Structured Prediction ◽  
Temporal Position ◽  
Audio Event ◽  
Weakly Supervised

Audio event detection (AED) is a task of recognizing the types of audio events in an audio stream and estimating their temporal positions. AED is typically based on fully supervised approaches, requiring strong labels including both the presence and temporal position of each audio event. However, fully supervised datasets are not easily available due to the heavy cost of human annotation. Recently, weakly supervised approaches for AED have been proposed, utilizing large scale datasets with weak labels including only the occurrence of events in recordings. In this work, we introduce a deep convolutional neural network (CNN) model called DSNet based on densely connected convolution networks (DenseNets) and squeeze-and-excitation networks (SENets) for weakly supervised training of AED. DSNet alleviates the vanishing-gradient problem and strengthens feature propagation and models interdependencies between channels. We also propose a structured prediction method for weakly supervised AED. We apply a recurrent neural network (RNN) based framework and a prediction smoothness cost function to consider long-term contextual information with reduced error propagation. In post-processing, conditional random fields (CRFs) are applied to take into account the dependency between segments and delineate the borders of audio events precisely. We evaluated our proposed models on the DCASE 2017 task 4 dataset and obtained state-of-the-art results on both audio tagging and event detection tasks.

Sign in / Sign up

Export Citation Format

Share Document