Scattering Invariants for Audio Classification
To obtain efficient feature representations for audio classification, it is desirable to have invariance to time-shift and stability to time-warping. Mel-frequency cepstral coefficients (MFCCs) satisfy these criteria, but are unsuitable for modeling large-scale temporal structure. The scattering transform extends this representation through a convolutional network of wavelet transforms and modulus operators, capturing structures at larger time scales. Additional invariance to frequency transposition with stability to frequency-warping is obtained by applying a second scattering transform along the log-frequency axis. Using these representations, we obtain state-of-the-art results on tasks such as phone segment classification and musical genre classification on the TIMIT and GTZAN datasets, respectively.
Видео Scattering Invariants for Audio Classification канала Microsoft Research
Видео Scattering Invariants for Audio Classification канала Microsoft Research
Показать
Комментарии отсутствуют
Информация о видео
Другие видео канала
![The Intern Experience at Microsoft Research Cambridge](https://i.ytimg.com/vi/qiCcJBZm8FY/default.jpg)
![Battling Tuberculosis Using Microsoft Technology](https://i.ytimg.com/vi/r2w40ZKxVUg/default.jpg)
![Fast and Flexible Multi-Task Classification Using Conditional Neural Adaptive Processes](https://i.ytimg.com/vi/apayUKSExmU/default.jpg)
![Data Visualization Reaches New Heights with Layerscape](https://i.ytimg.com/vi/Y4C2tvteQU4/default.jpg)
![Faculty Summit 2018 Introduction](https://i.ytimg.com/vi/RnzjxXOqovc/default.jpg)
![Microsoft Pix - Take Better Photos of People, Automatically](https://i.ytimg.com/vi/QXH_UaWMV48/default.jpg)
![Sirius: A Flat Datacenter Network with Nanosecond Optical Switching (SIGCOMM 2020 short talk)](https://i.ytimg.com/vi/y5Op0bhoLT8/default.jpg)
![Improvements on Higher Order Ambisonics Reproduction](https://i.ytimg.com/vi/DAlAaVTO8kg/default.jpg)
![Cambridge lab overview with Chris Bishop](https://i.ytimg.com/vi/8dR5lPav6K0/default.jpg)
![Get free cloud computing time and storage on Microsoft Azure](https://i.ytimg.com/vi/OLcOugO4Rqk/default.jpg)
![How interns at our New England Lab impact research at Microsoft](https://i.ytimg.com/vi/RNwIkn752uE/default.jpg)
![AI Institute "Geometry of Deep Learning" 2019 [Workshop] Day 1 | Session 2](https://i.ytimg.com/vi/FCmUtpkDk-I/default.jpg)
![De-Identifying Healthcare Data for Research](https://i.ytimg.com/vi/h-VhEVlC3h0/default.jpg)
![Intelligent cloud computing lifts villages out of water poverty](https://i.ytimg.com/vi/FCWMe1snMP8/default.jpg)
![How Microsoft and Novartis created Assess MS (short version)](https://i.ytimg.com/vi/R0dRuQ7Xy2w/default.jpg)
![Project Prague - The Producer - Behind the Scenes](https://i.ytimg.com/vi/UIakBZfEpPA/default.jpg)
![How OSIsoft and Deschutes Brewery used Microsoft Security Risk Detection](https://i.ytimg.com/vi/zMm3sUOm9jw/default.jpg)
![ChronoZoom curriculum and technology](https://i.ytimg.com/vi/awWCR81Wtdc/default.jpg)
![Recent Advances in Image Captioning, Image-Text Retrieval and…](https://i.ytimg.com/vi/4wS02nCWXvw/default.jpg)
![Using machine learning and AI to reduce hospital readmissions](https://i.ytimg.com/vi/bV8FHKCTx5k/default.jpg)
![IROS 2020 - Mixed Reality and Robotics Tutorial - Demo 1: Interaction](https://i.ytimg.com/vi/4G3wjPIs4Fc/default.jpg)