Sorry, you need to enable JavaScript to visit this website.

Content-Based Audio Processing

DEEP EMBEDDINGS FOR RARE AUDIO EVENT DETECTION WITH IMBALANCED DATA

Vipul Arora
Chao Wang

In this paper, we present a method to handle data imbalance for classification with neural networks, and apply it to acoustic event detection (AED) problem. The common approach to tackle data imbalance is to use class-weights in the objective function while training. An existing more sophisticated approach is to map the input to clusters in an embedding space, so that learning is locally balanced by incorporating inter-cluster and inter-class margins. On these lines, we propose a method to learn the embedding using a novel objective function, called triple-header cross entropy.

Paper Details

Authors:
Vipul Arora
Chao Wang
Vipul Arora, Ming Sun, Chao Wang
Submitted On:
10 May 2019 - 1:15am
Short Link:
Type:
Event:
Presenter's Name:
Paper Code:
Document Year:
Cite

Document Files

Poster for ICASSP2019

(314)

Keywords

Subscribe

Send Author a Private Message
[1] Vipul Arora, Ming Sun, Chao Wang, "DEEP EMBEDDINGS FOR RARE AUDIO EVENT DETECTION WITH IMBALANCED DATA", IEEE SigPort, 2019. [Online]. Available: http://sigport.org/4271. Accessed: Jun. 26, 2019.
@article{4271-19,
url = {http://sigport.org/4271},
author = {Vipul Arora; Ming Sun; Chao Wang },
publisher = {IEEE SigPort},
title = {DEEP EMBEDDINGS FOR RARE AUDIO EVENT DETECTION WITH IMBALANCED DATA},
year = {2019} }
TY - EJOUR
T1 - DEEP EMBEDDINGS FOR RARE AUDIO EVENT DETECTION WITH IMBALANCED DATA
AU - Vipul Arora; Ming Sun; Chao Wang
PY - 2019
PB - IEEE SigPort
UR - http://sigport.org/4271
ER -
Vipul Arora, Ming Sun, Chao Wang. (2019). DEEP EMBEDDINGS FOR RARE AUDIO EVENT DETECTION WITH IMBALANCED DATA. IEEE SigPort. http://sigport.org/4271
Vipul Arora, Ming Sun, Chao Wang, 2019. DEEP EMBEDDINGS FOR RARE AUDIO EVENT DETECTION WITH IMBALANCED DATA. Available at: http://sigport.org/4271.
Vipul Arora, Ming Sun, Chao Wang. (2019). "DEEP EMBEDDINGS FOR RARE AUDIO EVENT DETECTION WITH IMBALANCED DATA." Web.
1. Vipul Arora, Ming Sun, Chao Wang. DEEP EMBEDDINGS FOR RARE AUDIO EVENT DETECTION WITH IMBALANCED DATA [Internet]. IEEE SigPort; 2019. Available from : http://sigport.org/4271

Language Transfer of Audio Word2Vec: Learning Audio Segment Representations without Target Language Data

HUNG-YI LEE

Paper Details

Authors:
HUNG-YI LEE
Chia-Hao Shen, Janet Y. Sung, Hung-Yi Lee
Submitted On:
20 April 2018 - 12:59am
Short Link:
Type:
Event:
Document Year:
Cite

Document Files

Audio2vec (v7).pptx

(18)

Keywords

Subscribe

Send Author a Private Message
[1] Chia-Hao Shen, Janet Y. Sung, Hung-Yi Lee, "Language Transfer of Audio Word2Vec: Learning Audio Segment Representations without Target Language Data", IEEE SigPort, 2018. [Online]. Available: http://sigport.org/2866. Accessed: Jun. 26, 2019.
@article{2866-18,
url = {http://sigport.org/2866},
author = {Chia-Hao Shen; Janet Y. Sung; Hung-Yi Lee },
publisher = {IEEE SigPort},
title = {Language Transfer of Audio Word2Vec: Learning Audio Segment Representations without Target Language Data},
year = {2018} }
TY - EJOUR
T1 - Language Transfer of Audio Word2Vec: Learning Audio Segment Representations without Target Language Data
AU - Chia-Hao Shen; Janet Y. Sung; Hung-Yi Lee
PY - 2018
PB - IEEE SigPort
UR - http://sigport.org/2866
ER -
Chia-Hao Shen, Janet Y. Sung, Hung-Yi Lee. (2018). Language Transfer of Audio Word2Vec: Learning Audio Segment Representations without Target Language Data. IEEE SigPort. http://sigport.org/2866
Chia-Hao Shen, Janet Y. Sung, Hung-Yi Lee, 2018. Language Transfer of Audio Word2Vec: Learning Audio Segment Representations without Target Language Data. Available at: http://sigport.org/2866.
Chia-Hao Shen, Janet Y. Sung, Hung-Yi Lee. (2018). "Language Transfer of Audio Word2Vec: Learning Audio Segment Representations without Target Language Data." Web.
1. Chia-Hao Shen, Janet Y. Sung, Hung-Yi Lee. Language Transfer of Audio Word2Vec: Learning Audio Segment Representations without Target Language Data [Internet]. IEEE SigPort; 2018. Available from : http://sigport.org/2866

CONTENT-BASED REPRESENTATIONS OF AUDIO USING SIAMESE NEURAL NETWORKS

Pranay Manocha
Rohan Badlani
Anurag Kumar
Benjamin Elizalde
Ankit Shah

In this paper, we focus on the problem of content-based retrieval for
audio, which aims to retrieve all semantically similar audio recordings
for a given audio clip query. This problem is similar to the
problem of query by example of audio, which aims to retrieve media
samples from a database, which are similar to the user-provided example.
We propose a novel approach which encodes the audio into
a vector representation using Siamese Neural Networks. The goal is
to obtain an encoding similar for files belonging to the same audio

Paper Details

Authors:
Pranay Manocha
Rohan Badlani
Anurag Kumar
Benjamin Elizalde
Ankit Shah
Pranay Manocha, Rohan Badlani , Anurag Kumar , Ankit Shah, Benjamin Elizalde , Bhiksha Raj
Submitted On:
14 April 2018 - 3:50am
Short Link:
Type:
Event:
Presenter's Name:
Paper Code:
Document Year:
Cite

Document Files

ICASSP2018_Pranay.pdf

(183)

Keywords

Subscribe

Send Author a Private Message
[1] Pranay Manocha, Rohan Badlani , Anurag Kumar , Ankit Shah, Benjamin Elizalde , Bhiksha Raj, "CONTENT-BASED REPRESENTATIONS OF AUDIO USING SIAMESE NEURAL NETWORKS", IEEE SigPort, 2018. [Online]. Available: http://sigport.org/2810. Accessed: Jun. 26, 2019.
@article{2810-18,
url = {http://sigport.org/2810},
author = {Pranay Manocha; Rohan Badlani ; Anurag Kumar ; Ankit Shah; Benjamin Elizalde ; Bhiksha Raj },
publisher = {IEEE SigPort},
title = {CONTENT-BASED REPRESENTATIONS OF AUDIO USING SIAMESE NEURAL NETWORKS},
year = {2018} }
TY - EJOUR
T1 - CONTENT-BASED REPRESENTATIONS OF AUDIO USING SIAMESE NEURAL NETWORKS
AU - Pranay Manocha; Rohan Badlani ; Anurag Kumar ; Ankit Shah; Benjamin Elizalde ; Bhiksha Raj
PY - 2018
PB - IEEE SigPort
UR - http://sigport.org/2810
ER -
Pranay Manocha, Rohan Badlani , Anurag Kumar , Ankit Shah, Benjamin Elizalde , Bhiksha Raj. (2018). CONTENT-BASED REPRESENTATIONS OF AUDIO USING SIAMESE NEURAL NETWORKS. IEEE SigPort. http://sigport.org/2810
Pranay Manocha, Rohan Badlani , Anurag Kumar , Ankit Shah, Benjamin Elizalde , Bhiksha Raj, 2018. CONTENT-BASED REPRESENTATIONS OF AUDIO USING SIAMESE NEURAL NETWORKS. Available at: http://sigport.org/2810.
Pranay Manocha, Rohan Badlani , Anurag Kumar , Ankit Shah, Benjamin Elizalde , Bhiksha Raj. (2018). "CONTENT-BASED REPRESENTATIONS OF AUDIO USING SIAMESE NEURAL NETWORKS." Web.
1. Pranay Manocha, Rohan Badlani , Anurag Kumar , Ankit Shah, Benjamin Elizalde , Bhiksha Raj. CONTENT-BASED REPRESENTATIONS OF AUDIO USING SIAMESE NEURAL NETWORKS [Internet]. IEEE SigPort; 2018. Available from : http://sigport.org/2810

Knowledge Transfer From Weakly Labeled Audio Using Convolutional Neural Network For Sound Events and Scenes

Anurag Kumar

Paper Details

Authors:
Anurag Kumar
Anurag Kumar , Maksim Khadkevich , Christian Fugen
Submitted On:
14 April 2018 - 3:36am
Short Link:
Type:
Event:
Presenter's Name:
Paper Code:
Document Year:
Cite

Document Files

main.pdf

(186)

Keywords

Subscribe

Send Author a Private Message
[1] Anurag Kumar , Maksim Khadkevich , Christian Fugen, "Knowledge Transfer From Weakly Labeled Audio Using Convolutional Neural Network For Sound Events and Scenes", IEEE SigPort, 2018. [Online]. Available: http://sigport.org/2809. Accessed: Jun. 26, 2019.
@article{2809-18,
url = {http://sigport.org/2809},
author = {Anurag Kumar ; Maksim Khadkevich ; Christian Fugen },
publisher = {IEEE SigPort},
title = {Knowledge Transfer From Weakly Labeled Audio Using Convolutional Neural Network For Sound Events and Scenes},
year = {2018} }
TY - EJOUR
T1 - Knowledge Transfer From Weakly Labeled Audio Using Convolutional Neural Network For Sound Events and Scenes
AU - Anurag Kumar ; Maksim Khadkevich ; Christian Fugen
PY - 2018
PB - IEEE SigPort
UR - http://sigport.org/2809
ER -
Anurag Kumar , Maksim Khadkevich , Christian Fugen. (2018). Knowledge Transfer From Weakly Labeled Audio Using Convolutional Neural Network For Sound Events and Scenes. IEEE SigPort. http://sigport.org/2809
Anurag Kumar , Maksim Khadkevich , Christian Fugen, 2018. Knowledge Transfer From Weakly Labeled Audio Using Convolutional Neural Network For Sound Events and Scenes. Available at: http://sigport.org/2809.
Anurag Kumar , Maksim Khadkevich , Christian Fugen. (2018). "Knowledge Transfer From Weakly Labeled Audio Using Convolutional Neural Network For Sound Events and Scenes." Web.
1. Anurag Kumar , Maksim Khadkevich , Christian Fugen. Knowledge Transfer From Weakly Labeled Audio Using Convolutional Neural Network For Sound Events and Scenes [Internet]. IEEE SigPort; 2018. Available from : http://sigport.org/2809

CONTENT-BASED REPRESENTATIONS OF AUDIO USING SIAMESE NEURAL NETWORKS

Pranay Manocha
Anurag Kumar

In this paper, we focus on the problem of content-based retrieval for
audio, which aims to retrieve all semantically similar audio recordings
for a given audio clip query. This problem is similar to the
problem of query by example of audio, which aims to retrieve media
samples from a database, which are similar to the user-provided example.
We propose a novel approach which encodes the audio into
a vector representation using Siamese Neural Networks. The goal is
to obtain an encoding similar for files belonging to the same audio

Paper Details

Authors:
Pranay Manocha
Anurag Kumar
Pranay Manocha, Rohan Badlani, Anurag Kumar, Ankit Shah, Benjamin Elizalde, Bhiksha Raj
Submitted On:
14 April 2018 - 3:50am
Short Link:
Type:
Event:
Presenter's Name:
Paper Code:
Document Year:
Cite

Document Files

ICASSP2018_Pranay.pdf

(171)

Keywords

Subscribe

Send Author a Private Message
[1] Pranay Manocha, Rohan Badlani, Anurag Kumar, Ankit Shah, Benjamin Elizalde, Bhiksha Raj, "CONTENT-BASED REPRESENTATIONS OF AUDIO USING SIAMESE NEURAL NETWORKS", IEEE SigPort, 2018. [Online]. Available: http://sigport.org/2470. Accessed: Jun. 26, 2019.
@article{2470-18,
url = {http://sigport.org/2470},
author = {Pranay Manocha; Rohan Badlani; Anurag Kumar; Ankit Shah; Benjamin Elizalde; Bhiksha Raj },
publisher = {IEEE SigPort},
title = {CONTENT-BASED REPRESENTATIONS OF AUDIO USING SIAMESE NEURAL NETWORKS},
year = {2018} }
TY - EJOUR
T1 - CONTENT-BASED REPRESENTATIONS OF AUDIO USING SIAMESE NEURAL NETWORKS
AU - Pranay Manocha; Rohan Badlani; Anurag Kumar; Ankit Shah; Benjamin Elizalde; Bhiksha Raj
PY - 2018
PB - IEEE SigPort
UR - http://sigport.org/2470
ER -
Pranay Manocha, Rohan Badlani, Anurag Kumar, Ankit Shah, Benjamin Elizalde, Bhiksha Raj. (2018). CONTENT-BASED REPRESENTATIONS OF AUDIO USING SIAMESE NEURAL NETWORKS. IEEE SigPort. http://sigport.org/2470
Pranay Manocha, Rohan Badlani, Anurag Kumar, Ankit Shah, Benjamin Elizalde, Bhiksha Raj, 2018. CONTENT-BASED REPRESENTATIONS OF AUDIO USING SIAMESE NEURAL NETWORKS. Available at: http://sigport.org/2470.
Pranay Manocha, Rohan Badlani, Anurag Kumar, Ankit Shah, Benjamin Elizalde, Bhiksha Raj. (2018). "CONTENT-BASED REPRESENTATIONS OF AUDIO USING SIAMESE NEURAL NETWORKS." Web.
1. Pranay Manocha, Rohan Badlani, Anurag Kumar, Ankit Shah, Benjamin Elizalde, Bhiksha Raj. CONTENT-BASED REPRESENTATIONS OF AUDIO USING SIAMESE NEURAL NETWORKS [Internet]. IEEE SigPort; 2018. Available from : http://sigport.org/2470

AUTOMATIC MUSICAL KEY ESTIMATION WITH ADAPTIVE MODE BIAS

Gilberto Bernardes

In this paper we present the INESC Key Detection (IKD) system which incorporates a novel method for dynamically biasing key mode estimation using the spatial displacement of beat-synchronous Tonal Interval Vectors (TIVs). We evaluate the performance of the IKD system at finding the global key on three annotated audio datasets and using three key-defining profiles.

Paper Details

Authors:
Gilberto Bernardes
Gilberto Bernardes, Matthew E. P. Davies, and Carlos Guedes
Submitted On:
1 March 2017 - 2:13pm
Short Link:
Type:
Event:
Presenter's Name:
Paper Code:
Document Year:
Cite

Document Files

Poster_ICASSP.pdf

(1614)

Keywords

Subscribe

Send Author a Private Message
[1] Gilberto Bernardes, Matthew E. P. Davies, and Carlos Guedes, "AUTOMATIC MUSICAL KEY ESTIMATION WITH ADAPTIVE MODE BIAS", IEEE SigPort, 2017. [Online]. Available: http://sigport.org/1553. Accessed: Jun. 26, 2019.
@article{1553-17,
url = {http://sigport.org/1553},
author = {Gilberto Bernardes; Matthew E. P. Davies; and Carlos Guedes },
publisher = {IEEE SigPort},
title = {AUTOMATIC MUSICAL KEY ESTIMATION WITH ADAPTIVE MODE BIAS},
year = {2017} }
TY - EJOUR
T1 - AUTOMATIC MUSICAL KEY ESTIMATION WITH ADAPTIVE MODE BIAS
AU - Gilberto Bernardes; Matthew E. P. Davies; and Carlos Guedes
PY - 2017
PB - IEEE SigPort
UR - http://sigport.org/1553
ER -
Gilberto Bernardes, Matthew E. P. Davies, and Carlos Guedes. (2017). AUTOMATIC MUSICAL KEY ESTIMATION WITH ADAPTIVE MODE BIAS. IEEE SigPort. http://sigport.org/1553
Gilberto Bernardes, Matthew E. P. Davies, and Carlos Guedes, 2017. AUTOMATIC MUSICAL KEY ESTIMATION WITH ADAPTIVE MODE BIAS. Available at: http://sigport.org/1553.
Gilberto Bernardes, Matthew E. P. Davies, and Carlos Guedes. (2017). "AUTOMATIC MUSICAL KEY ESTIMATION WITH ADAPTIVE MODE BIAS." Web.
1. Gilberto Bernardes, Matthew E. P. Davies, and Carlos Guedes. AUTOMATIC MUSICAL KEY ESTIMATION WITH ADAPTIVE MODE BIAS [Internet]. IEEE SigPort; 2017. Available from : http://sigport.org/1553

COVER SONG IDENTIFICATION WITH 2D FOURIER TRANSFORM SEQUENCES

Prem Seetharaman

We approach cover song identification using a novel time-series representation of audio based on the 2DFT. The audio is represented as a sequence of magnitude 2D Fourier Transforms (2DFT). This representation is robust to key changes, timbral changes, and small local tempo deviations. We look at cross-similarity between these time-series, and extract a distance measure that is invariant to music structure changes. Our approach is state-of-the-art on a recent cover song dataset, and expands on previous work using the 2DFT for music representation and work on live song recognition.

Paper Details

Authors:
Prem Seetharaman
Prem Seetharaman, Zafar Rafii
Submitted On:
27 February 2017 - 3:00pm
Short Link:
Type:
Event:
Presenter's Name:
Paper Code:
Document Year:
Cite

Document Files

cover_song_poster.pdf

(367)

Keywords

Subscribe

Send Author a Private Message
[1] Prem Seetharaman, Zafar Rafii, "COVER SONG IDENTIFICATION WITH 2D FOURIER TRANSFORM SEQUENCES", IEEE SigPort, 2017. [Online]. Available: http://sigport.org/1445. Accessed: Jun. 26, 2019.
@article{1445-17,
url = {http://sigport.org/1445},
author = {Prem Seetharaman; Zafar Rafii },
publisher = {IEEE SigPort},
title = {COVER SONG IDENTIFICATION WITH 2D FOURIER TRANSFORM SEQUENCES},
year = {2017} }
TY - EJOUR
T1 - COVER SONG IDENTIFICATION WITH 2D FOURIER TRANSFORM SEQUENCES
AU - Prem Seetharaman; Zafar Rafii
PY - 2017
PB - IEEE SigPort
UR - http://sigport.org/1445
ER -
Prem Seetharaman, Zafar Rafii. (2017). COVER SONG IDENTIFICATION WITH 2D FOURIER TRANSFORM SEQUENCES. IEEE SigPort. http://sigport.org/1445
Prem Seetharaman, Zafar Rafii, 2017. COVER SONG IDENTIFICATION WITH 2D FOURIER TRANSFORM SEQUENCES. Available at: http://sigport.org/1445.
Prem Seetharaman, Zafar Rafii. (2017). "COVER SONG IDENTIFICATION WITH 2D FOURIER TRANSFORM SEQUENCES." Web.
1. Prem Seetharaman, Zafar Rafii. COVER SONG IDENTIFICATION WITH 2D FOURIER TRANSFORM SEQUENCES [Internet]. IEEE SigPort; 2017. Available from : http://sigport.org/1445

Conjugate gradient acceleration of non-linear smoothing filters

Andrew Knyazev

Noisy signal

The most efficient signal edge-preserving smoothing filters, e.g., for denoising, are non-linear. Thus, their acceleration is challenging and is often done in practice by tuning filters parameters, such as increasing the width of the local smoothing neighborhood, resulting in more aggressive smoothing of a single sweep at the cost of increased edge blurring. We propose an alternative technology, accelerating the original filters without tuning, by running them through a conjugate gradient method, not affecting their quality.

Paper Details

Authors:
Andrew Knyazev
Alexander Malyshev
Submitted On:
23 February 2016 - 1:44pm
Short Link:
Type:
Event:
Presenter's Name:
Document Year:
Cite

Document Files

GlobalSIP.pdf

(645)

Keywords

Subscribe

Send Author a Private Message
[1] Alexander Malyshev, "Conjugate gradient acceleration of non-linear smoothing filters", IEEE SigPort, 2015. [Online]. Available: http://sigport.org/406. Accessed: Jun. 26, 2019.
@article{406-15,
url = {http://sigport.org/406},
author = {Alexander Malyshev },
publisher = {IEEE SigPort},
title = {Conjugate gradient acceleration of non-linear smoothing filters},
year = {2015} }
TY - EJOUR
T1 - Conjugate gradient acceleration of non-linear smoothing filters
AU - Alexander Malyshev
PY - 2015
PB - IEEE SigPort
UR - http://sigport.org/406
ER -
Alexander Malyshev. (2015). Conjugate gradient acceleration of non-linear smoothing filters. IEEE SigPort. http://sigport.org/406
Alexander Malyshev, 2015. Conjugate gradient acceleration of non-linear smoothing filters. Available at: http://sigport.org/406.
Alexander Malyshev. (2015). "Conjugate gradient acceleration of non-linear smoothing filters." Web.
1. Alexander Malyshev. Conjugate gradient acceleration of non-linear smoothing filters [Internet]. IEEE SigPort; 2015. Available from : http://sigport.org/406