Filters








2,900 Hits in 6.4 sec

Semi-supervised Acoustic Event Detection based on tri-training [article]

Bowen Shi, Ming Sun, Chieh-Chi Kao, Viktor Rozgic, Spyros Matsoukas, Chao Wang
2019 arXiv   pre-print
Based on the classic tri-training approach, our proposed method shows accuracy improvement over both the supervised training baseline, and semisupervised self-training set-up, in all pre-defined acoustic  ...  Recent acoustic event detectors are based on large-scale neural networks, which are typically trained with huge amounts of labeled data.  ...  We propose an ensemble method based on the classic tri-training that shows improvements in all acoustic events we investigate in a realistic semi-supervised setting (Internet-scale unlabeled dataset with  ... 
arXiv:1904.12926v1 fatcat:sjx4qz4b3vccvp7ocwgn4nlvp4

Semi-supervised Acoustic Event Detection Based on Tri-training

Bowen Shi, Ming Sun, Chieh-Chi Kao, Viktor Rozgic, Spyros Matsoukas, Chao Wang
2019 ICASSP 2019 - 2019 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)  
Based on the classic tri-training approach, our proposed method shows accuracy improvement over both the supervised training baseline, and semisupervised self-training set-up, in all pre-defined acoustic  ...  Recent acoustic event detectors are based on large-scale neural networks, which are typically trained with huge amounts of labeled data.  ...  We propose an ensemble method based on the classic tri-training that shows improvements in all acoustic events we investigate in a realistic semi-supervised setting (Internet-scale unlabeled dataset with  ... 
doi:10.1109/icassp.2019.8683710 dblp:conf/icassp/ShiSKRMW19 fatcat:33op2g5rtbavjfjlyrjee5hziu

Audio Event and Scene Recognition: A Unified Approach using Strongly and Weakly Labeled Data [article]

Anurag Kumar, Bhiksha Raj
2017 arXiv   pre-print
The primary problem domain focus of this paper is acoustic event and scene detection in audio recordings. We first propose a naive formulation for leveraging labeled data in both forms.  ...  We then propose a more general framework for Supervised and Weakly Supervised Learning (SWSL). Based on this general framework, we propose a graph based approach for SWSL.  ...  EXPERIMENTS AND RESULTS We evaluate the proposed supervised and weakly supervised learning on both audio event and acoustic scene detection tasks.  ... 
arXiv:1611.04871v3 fatcat:jclmmvkhzrhyfgjygjpxxafw6a

Peer Collaborative Learning for Polyphonic Sound Event Detection [article]

Hayato Endo, Hiromitsu Nishizaki
2021 arXiv   pre-print
This paper describes that semi-supervised learning called peer collaborative learning (PCL) can be applied to the polyphonic sound event detection (PSED) task, which is one of the tasks in the Detection  ...  and Classification of Acoustic Scenes and Events (DCASE) challenge.  ...  This paper adopts a knowledge distillation model based on semi-supervised learning called "peer collaborative learning" (PCL) [16] .  ... 
arXiv:2110.03511v1 fatcat:cbor4tqn2vadxbb3hrwbclfhle

Weakly Labeled Sound Event Detection Using Tri-training and Adversarial Learning [article]

Hyoungwoo Park, Sungrack Yun, Jungyun Eum, Janghoon Cho, Kyuwoong Hwang
2019 arXiv   pre-print
This paper considers a semi-supervised learning framework for weakly labeled polyphonic sound event detection problems for the DCASE 2019 challenge's task4 by combining both the tri-training and adversarial  ...  Given this dataset, we apply the tri-training where two different classifiers are used to obtain pseudo labels on the weakly labeled and unlabeled dataset, and the final classifier is trained using the  ...  This paper presents a sound event detection combining adversarial learning and tri-training.  ... 
arXiv:1910.06790v1 fatcat:x5kcng4kcncgfc3odugzmcghbm

An Approach for Self-Training Audio Event Detectors Using Web Data [article]

Benjamin Elizalde, Ankit Shah, Siddharth Dalmia, Min Hun Lee, Rohan Badlani, Anurag Kumar, Bhiksha Raj, Ian Lane
2017 arXiv   pre-print
Audio Event Detection (AED) aims to recognize sounds within audio and video recordings. AED employs machine learning algorithms commonly trained and tested on annotated datasets.  ...  The audio event detectors are trained on the labeled audio and ran on the unlabeled audio downloaded from YouTube.  ...  Unlabeled audio from videos can help audio event detection. arXiv:1609.06026v3 [cs.SD] 27 Jun 2017 II. SEMI-SUPERVISED SELF-TRAINING OF AUDIO EVENT DETECTORS Fig. 1 . 1 Fig. 1.  ... 
arXiv:1609.06026v3 fatcat:cn4z3paieffphntfmt7ww756qy

Polyphonic Music Note Onset Detection Using Semi-Supervised Learning

Wei You, Roger B. Dannenberg
2007 Zenodo  
We also tried training on the score-alignment data without bootstrapping. The results show it is important to have both a large data set and actual acoustic data.  ...  Semi-supervised learning From score alignment, we obtain a large set of labeled training data, but the labels are based on rather large windows, and the chroma vector features are chosen more for gross  ... 
doi:10.5281/zenodo.1417384 fatcat:p5d27s5tlnchjkpsyyfj2uvfpa

An Approach For Self-Training Audio Event Detectors Using Web Data

Rohan Badlani, Siddharth Dalmia, Benjamin Elizalde, Anurag Kumar, Ian Lane, Min Hun Lee, Bhiksha Raj, Ankit Shah
2018 Zenodo  
Unlabeled audio from videos can help audio event detection. Fig. 1 . 1 Flow of the semi-supervised self-training of Audio Event Detection.  ...  SEMI-SUPERVISED SELF-TRAINING OF AUDIO EVENT DETECTORS Semi-supervised self-training is an algorithm that iteratively re-trains a model and our particular framework is illustrated in Figure 1 .  ... 
doi:10.5281/zenodo.1159183 fatcat:k6kuonmwlrckzjwurzwomdpsg4

Detecting Gunshots Using Wearable Accelerometers

Charles E. Loeffler, Stefan Strack
2014 PLoS ONE  
Existing monitoring systems rely heavily on location-based monitoring methods, which have incomplete geographic coverage and do not provide information on illegal firearm use.  ...  These results suggest the feasibility of using inexpensive wearable sensors to detect firearm discharges.  ...  Instead, a motion-based detection framework was chosen.  ... 
doi:10.1371/journal.pone.0106664 pmid:25184416 pmcid:PMC4153670 fatcat:p4q746w4xbg57bcom7d4shi554

Acoustic anomaly detection via latent regularized gaussian mixture generative adversarial networks [article]

Chengwei Chen and Pan Chen and Lingyu Yang and Jinyuan Mo and Haichuan Song and Yuan Xie and Lizhuang Ma
2020 arXiv   pre-print
Acoustic anomaly detection aims at distinguishing abnormal acoustic signals from the normal ones. It suffers from the class imbalance issue and the lacking in the abnormal instances.  ...  In this paper, a novel Gaussian Mixture Generative Adversarial Network (GMGAN) is proposed under semi-supervised learning framework, in which the underlying structure of training data is not only captured  ...  anomaly detection to learn an regularized latent space by using adversarial training under semi-supervised learning framework.  ... 
arXiv:2002.01107v2 fatcat:r4saotwkrrhmzay3rcscrqwwqa

Semi-Supervised Deep Learning in High-Speed Railway Track Detection Based on Distributed Fiber Acoustic Sensing

Shulun Wang, Feng Liu, Bin Liu
2022 Sensors  
We propose a track detection method that innovatively leverages semi-supervised deep learning based on image recognition, with a particular pre-processing for the dataset and a greedy algorithm for the  ...  Therefore, approaches based on optical sensors have become the most remarkable strategy in terms of deployment cost and real-time performance.  ...  In this paper, we propose a track detection system that innovatively leverages semi-supervised deep learning based on image recognition.  ... 
doi:10.3390/s22020413 pmid:35062373 pmcid:PMC8779117 fatcat:k7lstxzn3jhmhozgefjmhlrwo4

Review of fall detection techniques: A data availability perspective

Shehroz S. Khan, Jesse Hoey
2017 Medical Engineering and Physics  
The taxonomy identifies different categories of classification methods for the study of fall detection based on the availability of their data during training the classifiers.  ...  Therefore, standard supervised machine learning methods may not be directly applied to handle this problem.  ...  Fahmi et al. [59] present a semi-supervised fall detection method using smartphones by first training a supervised algorithm using decision trees, then using fall profiles to develop a semi-supervised  ... 
doi:10.1016/j.medengphy.2016.10.014 pmid:27889391 fatcat:pdpd5om4pjfqhmavxaqxkogjz4

An Analysis of Sound Event Detection under Acoustic Degradation Using Multi-Resolution Systems

Diego de Benito-Gorrón, Daniel Ramos, Doroteo T. Toledano
2021 Applied Sciences  
The Sound Event Detection task aims to determine the temporal locations of acoustic events in audio clips.  ...  the DCASE Challenge (Detection and Classification of Acoustic Scenes and Events).  ...  Conformer-Based Sound Event Detection with Semi-Supervised Learning and Data Augmentation.  ... 
doi:10.3390/app112311561 fatcat:eejcrjkkxvfwlkrppl7xztsbiu

Semi-supervised Triplet Loss Based Learning of Ambient Audio Embeddings

Nicolas Turpault, Romain Serizel, Emmanuel Vincent
2019 ICASSP 2019 - 2019 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)  
In this paper, we combine unsupervised and supervised triplet loss based learning into a semi-supervised representation learning approach.  ...  They have compared this approach to supervised learning. However, in real situations, it is common to have a small labeled dataset and a large unlabeled one.  ...  All classifiers are trained on the 1,578 labeled files from DCASE 2018 task 4 training set.  ... 
doi:10.1109/icassp.2019.8683774 dblp:conf/icassp/TurpaultSV19 fatcat:f4pcyw36gbh2fd4sqwxi3jkkem

Open Set Audio Classification Using Autoencoders Trained on Few Data

Javier Naranjo-Alcazar, Sergi Perez-Castanos, Pedro Zuccarello, Fabio Antonacci, Maximo Cobos
2020 Sensors  
on latent space representations to detect known classes and reject unwanted ones.  ...  on transfer learning.  ...  Results(%) of the proposed frameworks using DCASE (Detection and Classification of Acoustic Scenes and Events) 2019 Task 1C dataset.  ... 
doi:10.3390/s20133741 pmid:32635378 pmcid:PMC7374438 fatcat:nl3iuuijpjhnxic6nyvkgm57ba
« Previous Showing results 1 — 15 out of 2,900 results