Domenico Stefani

ITIS "De Pretto" Schio

Università di Trento

Completion/conferral date: 11 Jan 2024Thesis: Embedded Real-time Deep Learning for a Smart Guitar: A Case Study on Expressive GuitarTechnique RecognitionTechnical skills developed:• Audio Plugin development with C++ and the JUCE framework,• Audio signal processing and feature extraction,• Deployment of plugins to Embedded devices (RPi 4) with Elk Audio OS,• Tensorflow + Python programming for Music Information Retrieval,• Neural network inference in C++… Show more Completion/conferral date: 11 Jan 2024Thesis: Embedded Real-time Deep Learning for a Smart Guitar: A Case Study on Expressive GuitarTechnique RecognitionTechnical skills developed:• Audio Plugin development with C++ and the JUCE framework,• Audio signal processing and feature extraction,• Deployment of plugins to Embedded devices (RPi 4) with Elk Audio OS,• Tensorflow + Python programming for Music Information Retrieval,• Neural network inference in C++ (TensorFlow Lite, Libtorch/Torchscript, ONNX Runtime, RTNeural).• Academic Writing (English, Two C1 Courses attended and passed with full marks). Show less

Università di Trento

Thesis: Embedded real-time classification of percussive and pitched sounds on a smart guitarMy free-choice credits were allocated with particular interest to courses such as Multisensory Interactive Systems, Machine Learning and Deep Learning.A relevant project, developed for the Deep Learning course, revolved around objective music evaluation and automatic generation of music scores using machine learning.Thesis topic involved real time feature extraction and… Show more Thesis: Embedded real-time classification of percussive and pitched sounds on a smart guitarMy free-choice credits were allocated with particular interest to courses such as Multisensory Interactive Systems, Machine Learning and Deep Learning.A relevant project, developed for the Deep Learning course, revolved around objective music evaluation and automatic generation of music scores using machine learning.Thesis topic involved real time feature extraction and classification of audio signals and C++ developement for audio DSP. Show less

Università degli Studi di Trento

Thesis developed on a company project. The project itself was developed for Vivica S.r.l. and it consists in a system that allows to connect to proprietary devices through UDP messages. The systems included the developement of a C++ Client Library, a C++ Server library for the proprietary devices and an Android app with Native support for the C++ library. Everthing was developed with cross-compilation and cross-platform support in mind.

Publicity Chair and Virtual Zoom meeting host

Associated with Domenico Stefani

Volontario

Associated with Domenico Stefani