CaR-FOREST: Joint Classification-Regression Decision Forests for Overlapping Audio Event Detection

This report describes our submissions to Task2 and Task3 of the DCASE 2016 challenge. The systems aim at dealing with the detection of overlapping audio events in continuous streams, where the detectors are based on random decision forests. The proposed forests are jointly trained for classification and regression simultaneously. Initially, the training is classification-oriented to encourage the trees to select discriminative features from overlapping mixtures to separate positive audio segments from the negative ones. The regression phase is then carried out to let the positive audio segments vote for the event onsets and offsets, and therefore model the temporal structure of audio events. One random decision forest is specifically trained for each event category of interest. Experimental results on the development data show that our systems significantly outperform the baseline on the Task2 evaluation while they are inferior to the baseline in the Task3 evaluation.

• Publication year

  2016

• Venue

  arXiv.org

• Publication date

  2016-07-08

• Fields of study

  Computer Science

• Identifiers
  arXiv 1607.02306
• External record

  Open on Semantic Scholar

• Source metadata

  Semantic Scholar

• No claims are published for this paper.

• No concepts are published for this paper.

• TUT database for acoustic scene classification and sound event detection

  2016cited by this paper
• Learning compact structural representations for audio events using regressor banks

  2016cited by this paper
• Early event detection in audio streams

  2015cited by this paper
• A multi-channel fusion framework for audio event detection

  2015cited by this paper
• Random Regression Forests for Acoustic Event Detection and Classification

  2015influential reference
• Temporal Acoustic Words for Online Acoustic Event Detection

  2015cited by this paper
• Polyphonic sound event detection using multi label deep neural networks

  2015cited by this paper
• A Bag-of-Features approach to acoustic event detection

  2014cited by this paper
• Acoustic event detection and localization with regression forests

  2014influential reference
• Real-Time Detection of Overlapping Sound Events with Non-Negative Matrix Factorization

  2013cited by this paper
• Detection and Classification of Acoustic Events for In-Home Care

  2011cited by this paper
• Sound Event Detection in Multisource Environments Using Source Separation

  2011cited by this paper
• Jump Function Kolmogorov for overlapping audio event classification

  2011cited by this paper
• Machine Hearing: An Emerging Field

  2010cited by this paper
• Real-world acoustic event detection

  2010cited by this paper
• Acoustic event detection in real life recordings

  2010cited by this paper
• Acoustic event detection in meeting-room environments

  2009cited by this paper
• Computational Auditory Scene Analysis: Principles, Algorithms, and Applications

  2008cited by this paper
• Scream and gunshot detection and localization for audio-surveillance systems

  2007cited by this paper
• Noise power spectral density estimation based on optimal smoothing and minimum statistics

  2001cited by this paper

• Simultaneous regression and classification for drug sensitivity prediction using an advanced random forest method

  2022cites this paper
• Fe b 20 19 UNIFYING ISOLATED AND OVERLAPPING AUDIO EVENT DETECTION WITH MULTI-LABEL MULTITASK CONVOLUTIONAL RECURRENT NEURAL NETWORKS

  2021cites this paper
• Detecting Overlapping Data in System Logs Based on Ensemble Learning Method

  2020cites this paper
• Multi-Task Learning for Acoustic Event Detection Using Event and Frame Position Information

  2020influential citation
• Acoustic event detection utilizing event class and localization information

  2019influential citation
• Proceedings of the Detection and Classification of Acoustic Scenes and Events 2018 Workshop (DCASE2018)

  2018cites this paper
• Acoustic Features for Environmental Sound Analysis

  2018cites this paper
• Weighted and Multi-Task Loss for Rare Audio Event Detection

  2018cites this paper
• Convolutional Neural Networks with Multi-task Loss for Polyphonic Sound Event Detection

  2018cites this paper
• Unifying Isolated and Overlapping Audio Event Detection with Multi-label Multi-task Convolutional Recurrent Neural Networks

  2018cites this paper
• Classifier Architectures for Acoustic Scenes and Events: Implications for DNNs, TDNNs, and Perceptual Features from DCASE 2016

  2017cites this paper
• DNN and CNN with Weighted and Multi-task Loss Functions for Audio Event Detection

  2017cites this paper
• Chapter 4 : Acoustic Features for Environmental Sound Analysis

  2017cites this paper
• Enabling Early Audio Event Detection with Neural Networks

  2017cites this paper
• SOUND EVENT DETECTION USING DEEP RANDOM FOREST

  2017cites this paper
• Sound event detection in synthetic audio: Analysis of the dcase 2016 task results

  2017influential citation
• and Multi-Task Loss for Rare

  year unknowncites this paper
• Isolated and Overlapping Audio Event Detection with Multi-Label Multi-Task Convolutional Recurrent Neural Networks

  year unknowncites this paper