|
Fernando de la Calle Silos 
Researcher and Associate Professor fsilos [at] tsc [dot] uc3m [dot] es
|
|
|
Introduction |
|
|
During my years as a researcher I have developed a deep passion for signal processing and machine learning, focusing on diverse areas such as Speech Recognition and Computer Vision. I am very passionate to be working on these topics as part of my research at University Carlos III of Madrid and Carnegie Mellon University, and would like to keep doing research on related areas. Detailed information can be found on my curriculum vitae. |
|
|
Publications |
|
|
Dissertation |
Bio-Motivated Features and Deep Learning for Robust Speech Recognition |
|
Journal |
Synchrony-Based Feature Extraction for Robust Automatic Speech Recognition |
|
Morphologically- filtered power-normalized cochleograms as robust, biologically inspired features for ASR |
|
|
International Conferences |
Deep Residual Networks with Auditory Inspired Features for Robust Speech Recognition |
|
ASR Feature Extraction with Morphologically-Filtered Power-Normalized Cochleograms |
|
|
Mid-Level Feature Set for Specific Event and Anomaly Detection in Crowded Scenes. |
|
|
Local Conferences |
An Analysis of Deep Neural Networks in Broad Phonetic Classes for Noisy Speech Recognition |
|
Preliminary experiments on the robustness of biologically motivated features for DNN-based ASR |
|
|
Deep Maxout Networks applied to Noise-Robust Speech Recognition |
|
|
Education |
|
|
Phd |
Bio-Motivated Features and Deep Learning for Robust Speech Recognition |
|
MSc |
Master in Multimedia and Communications. |
|
BSc |
Bachelor in Telecommunication Technology Engineering |
|
Teaching |
|
|
Speech, Audio, Image, and Video Processing Applications |
|
|
Algorithms for Multimedia Information Management |
|
|
Multimedia Information Coding in Communications |
|
|
Acoustical Instrumentation and Noise Control |
|
|
Digital Audio Processing |
|
|
Code |
|
|
fsilosSpeechToolbox: Implementation of all the feature extraction methods presented in my PhD thesis. ResNet-Kaldi-Tensorflow-ASR: ResNet and other CNN implementations in Tensorflow presented in the paper: Deep Residual Networks with Auditory Inspired Features for Robust Speech Recognition. MI_Feature_Selection: Matlab code that implements the feature selection algorithm using mutual information described in the ICIP 2013 paper |
|
