Computer Science Department propose a PhD dissertation

Computer Science Department at UOT discussed a Ph.D. dissertation about "the discovery of the speaker's gender and estimation of its age for the interaction between humans and robots"

The dissertation of the student Amir Abdul-Baqi Badr aimed to propose a new automated system for discovering the speaker’s gender and estimating his age in (HRI) environment, for the interaction between humans and robots so as to bridge the gap in choosing the optimal set of features by choosing a set of features that effectively addresses the problem of discovering the speaker’s gender estimate its life and at the same time reduce the response time to keep the robot's reaction time as short as possible.

The researcher showed that predicting gender and age information from short utterance has become vital for many applications in daily life, such as targeted marketing, human-robot interaction (HRI), and identification of suspects in criminal cases, among others. Despite extensive experiments conducted to estimate such data, researchers still have to develop an efficient system for such tasks in an (HRI) environment.

The researcher concluded that in detecting the gender of the speaker, the detection errors from the SSC, Fractals and F0 feature sets are complementary, allowing these features to be combined to further enhance the results. The use of the automatic encoding method for feature selection has a significant impact on the performance of the speaker age estimation system. CatBoost's feature selection method improves the performance of the speaker age and gender recognition system.

The discussion committee consisted of: -

- Prof. Dr. Bilal Ismail from Al-Anbar University/ the head.

- Prof. Dr. Khaled Ali from Al-Mustansiriya University / member.

- Assist. Prof. Dr. Amer Siddiq from Imam Jaafar Al-Sadiq University / member.

- Assist. Prof. Dr. Nida'a Falih from the University of Technology / member.

 - Assist. Prof. Dr. Yousra Hussein from the University of Technology / member.