TUT database for acoustic scene classification and sound event detection

By Annamaria Mesaros, Toni Heittola, Tuomas Virtanen. 2016

660 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition, Signal Processing

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

In Signal Processing

Citations in the corpus

Speech Denoising with Deep Feature Losses (2018, methodology): “For the scene classification task, the training set [37] consists of 30-second-long audio files sampled at 44.”
FSD50K: An Open Dataset of Human-Labeled Sound Events (2020, background): “In contrast, previous SER datasets were of a more limited size and coverage (e.g. [21]– [23]).”
CNN architectures for large-scale audio classification (2016, methodology): “Prior work has been reported on datasets such as TRECVid [14], ActivityNet [15], Sports1M [16], and TUT/DCASE Acoustic scenes 2016 [17] which are much smaller than YouTube-100M.”
PANNs: Large-Scale Pretrained Audio Neural Networks for Audio Pattern Recognition (2019, methodology): “pretrained on MagnaTagATune [23] and acoustic scene [24] datasets were fine-tuned on other audio tagging tasks [25], [26].”
SONYC Urban Sound Tagging (SONYC-UST): A Multilabel Dataset from an Urban Acoustic Sensor Network (2019, background): “The TUT Sound Events 2016 [10, 11, 12] and 2017 [13, 14] datsets consists of audio recordings in real urban environments as well as providing strong, multilabel annotations.”
A Closer Look at Weak Label Learning for Audio Events (2018, background): “datasets and challenges released in the last decade or so [2], [3], [4], [5], [6].”
Active Learning for Sound Event Detection (2020, methodology): “log-mel spectrogram in an environmental sound classification dataset [15] and an acoustic scene classification dataset [19].”
What’s all the Fuss about Free Universal Sound Separation Data? (2020, methodology): “Approaches have been developed to deal with overlapping target sound events using multilabeled (foreground vs background) training sets [10], sets of binary classiﬁers [11], factorization techniques on the input of the classiﬁer [12, 13], or exploiting spatial information when available [14].”
Audio event and scene recognition: A unified approach using strongly and weakly labeled data (2016, methodology): “We work with a toal of 15 acoustic scenes from DCASE [14] dataset, which is also source of our strongly labeled data.”
The Benefit of Temporally-Strong Labels in Audio Event Classification (2021, background): “For example, the TUT Sound events 2016 [3] and TUT Sound events 2017 [4] each total ∼2 h of labeled audio.”
Freesound Datasets: A Platform for the Creation of Open Audio Datasets (2017, background): “5 • TUT Acoustic Scenes [23] is a publicly available dataset released for the DCASE 2016 challenge.”
DCASE2017 Challenge Setup: Tasks, Datasets and Baseline System (2017, methodology): “A detailed description of the data recording and annotation procedure can be found in [26].”
Multi-label Few-shot Learning for Sound Event Recognition (2019, background): “The task has essential applications in many fields such as surveillance, smart city, and smart cars [1], [2].”
The SINS Database for Detection of Daily Activities in a Home Environment Using an Acoustic Sensor Network (2017, methodology): “The LITIS Rouen Audio Scene dataset [12], DCASE 2013 and 2016 databases [13, 14] consist of (binaural) recordings of events and/or scenes in public areas, e.”
AudioCaps: Generating Captions for Audios in The Wild (2019, background): “There exist a number of datasets to aid in achieving this goal, including DCASE series (Stowell et al., 2015; Mesaros et al., 2016, 2017), UrbanSound8k (Salamon et al., 2014), ESC (Piczak, 2015).”
VOICe: A Sound Event Detection Dataset For Generalizable Domain Adaptation (2019, methodology): “For example, in the DCASE workshop and challenge2, the recent state-of-the-art results are obtained by applying stacked convolutional and recurrent neural networks (CRNNs) [2, 8].”
A Deep Residual Network for Large-Scale Acoustic Scene Analysis (2019, background): “There has been a lot of progress thanks to many publicly available datasets, such as ESC, those in the DCASE challenges, and AudioSet [7, 8, 9, 10].”
Vggsound: A Large-Scale Audio-Visual Dataset (2020, background): “However in the audio domain, while several audio datasets have been released in the past few years [6, 7, 8, 9], the data collection process usually requires extensive human efforts, making it unscalable and often limited to narrow domains.”
Audio Set Classification with Attention Model: A Probabilistic Perspective (2017, methodology): “In DCASE challenge, several tasks have been defined for audio classification including acoustic scene classification [1], sound event detection [1] and audio tagging [3, 4].”