Emotional state and acute diseases can be predicted with the software developed by a student and research assistant at Kırıkkale University, Faculty of Engineering and Natural Sciences, Department of Computer Engineering.
The project titled “Gender Prediction from Voice Signals with Multiple Feature Fusion”, prepared by Research Assistant Mustafa Yurdakul and 3rd grade student Açelya Gümele, received support for 1 year within the scope of TÜBİTAK’s “University Students Research Projects Support Program”.
Project consultant Mustafa Yurdakul said that they developed software with Gümele within the scope of TÜBİTAK’s program.
Stating that they aimed to perform gender analysis from voice with the software, Yurdakul said, “When we got good results from the technique we developed, we used the application in various fields. We also achieved successful results in areas such as mood and disease prediction from voice. We obtained a model with high performance and accuracy.” said.
Yurdakul gave information about the software and stated that they first record the user’s voice for 3-5 seconds.
Explaining that they then analyzed the sound with various techniques, Yurdakul said:
“In this way, we can predict various factors from the voice. The technique we proposed to TUBITAK was only about estimating gender from the voice. When the results we obtained were very good, we decided to use it in the field of health as well. We give the short-term sounds we receive from the patients as input to our model, and we also include certain acute or chronic diseases. We can guess.”
Yurdakul stated that the software can help healthcare professionals with the diagnosis of diseases and security forces with voice analysis of the suspect during the statement.
Açelya Gümele, a 3rd year student at the computer department, stated that they aim to create a gender recognition system from voice data with artificial intelligence in the project, and said, “It can be used in creating an extra layer of security in phone verification and online transactions, and in the diagnosis and treatment of voice disorders in the health sector. Our project has commercial potential and is available for marketing purposes.” “From this perspective, it will also help companies personalize and target their strategies.” he said.
1355
