| Grant number: | 18/26455-8 |
| Support Opportunities: | Regular Research Grants |
| Start date: | May 01, 2019 |
| End date: | October 31, 2022 |
| Field of knowledge: | Engineering - Electrical Engineering - Telecommunications |
| Principal Investigator: | Miguel Arjona Ramírez |
| Grantee: | Miguel Arjona Ramírez |
| Host Institution: | Escola Politécnica (EP). Universidade de São Paulo (USP). São Paulo , SP, Brazil |
| City of the host institution: | São Paulo |
| Associated researchers: | Celso Setsuo Kurashima ; Demostenes Zegarra Rodriguez ; Edmilson da Silva Morais ; Emilio Del Moral Hernandez ; Mario Minami ; Renata Lopes Rosa ; Wesley Beccaro |
| Associated scholarship(s): | 20/13459-5 - Evaluation of voice quality using machine learning and considering wireless networks,
BP.TT 19/10329-6 - Evaluation of voice quality using machine learning and considering wireless networks, BP.TT 19/10551-0 - Video quality assessment using machine learning in network scenarios, BP.TT |
Abstract
This research plan addresses a common basis for a number of areas in signal processing such as speech analysis, speech coding and audio coding, speech recognition and audio feature recognition as well as source separation with regularizations to carry out adjustments suitable to the desired application. Traditionally, speech analysis, in addition to its own importance, also provides signal representations and model parameters that are necessary to the other areas. In this role it is losing appeal with deep learning and parallels are set to be established in order to bring about some interpretation. Beyond usual types of time-frequency decomposition and modification and autoregressive analysis, new algorithms will be explored and proposed based on machine learning and deep learning for enhancement, separation and synthesis of speech and audio signals, partially or totally replacing traditional analysis. Research will focus on generative machines capable of handling video signals and time series as well.Additionally, the parameters and representations of the speech signal will also be used to model and elaborate non-intrusive speech quality metrics; for this purpose, the speech signal is degraded using different communication system parameters. (AU)
| Articles published in Agência FAPESP Newsletter about the research grant: |
| More itemsLess items |
| TITULO |
| Articles published in other media outlets ( ): |
| More itemsLess items |
| VEICULO: TITULO (DATA) |
| VEICULO: TITULO (DATA) |