[contact-form-7 404 "Not Found"]

Case Studies

Who Speaks

Zesium is a company with more than 70 projects successfully implemented in various languages, fields, and countries. As one of the partners of The Faculty of Technical Sciences at The University of Novi Sad, we gather the most brilliant minds who work on the research programs for our company including machine learning, voice recognition, signal processing, neural networks and data mining.

Who Speaks is a research project of highly experienced engineers, brilliant developers and skillful management. Moreover, previous successfully implemented projects in Israel (where ‘Who Speaks’ clients are from) were the reason for our partners to believe we are familiar with Israeli way of working and reliable for the project to be a high quality one and delivered on time.

Who Speaks is an innovative application which allows users to analyze voice frequencies and recognize the number of speakers/voices in a room – we entered a new research projects related to the voice recognition and voice processing. The requirements were to have a possibility to register as many voice frequencies in one place as possible and to collect data about these voices/frequencies all in one application. Zesium created an innovative algorithm able to recognize voices at one place and make the communication between Android UI and backend side (RESTful services) possible. Application is able to recognize how many people are there in the room, based on their voice frequencies, and subsequently recognize and mark each speaker at the place. The created User Interface is able to name different speakers in the room, distinguish them and mark their voice each time they speak.

Who Speaks mobile app

The Who Speaks application is able to analyze and segment specific sequences directly from the raw data based on certain parameters from the algorithm. The challenge of the app development was to show graphical redistribution of specific speakers who participated in the conversation and to recognize their involvement and voices in real time. Technologies used in the development process of the application are: C++ and Android Studio.

The Who Speaks application has evolved along the time and has incredible potential to be developed further in the future, as speech recognition is a hot topic not just in IT world, but also in other fields such as neural networks, healthcare, education, military and everyday life.

This was a significant project for Zesium, as it gave an opportunity to our developers and scientists to expand their knowledge further in the field of voice recognition and to successfully finish Who Speaks project.


You have a question?
Book a free 60 min consultation!