Facing stiff competition, Mr. Tran Bao Toan, from the DTU Center of Software Engineering, won first prize in the Emotion Recognition Challenge (ERC2019) contest with his “Emotion Recognition Machine Learning from Voice” application.
Mr. Tran Bao Toan (center)
The ERC2019 topic, entitled “Pioneering solutions for Emotion recognition”, concentrated on a variety of text, image, sound and video data types and this was the first contest based on sound-based emotion recognition to be held in Vietnam and the contest identifies and recognizes talented individuals from Vietnam and elsewhere.
Entrants first receive a training dataset containing 7,442 samples organized in wave-label file pairs as follows:
- Wave: spoken data in PCM 16 kHz, 16 bit, mono format
- Label: a text file containing six emotion tag wave files for: happiness, sadness, anger, fear, disgust and neutrality.
Winning teams and officials
Reporter: How did you develop your application from the training dataset?
Mr. Tran Bao Toan: I created the machine learning model using an artificial neural network, which was able to distinguish between the six human emotions through voice recordings. I then uploaded my model to compete with the other contestants. The most accurate nine models were the selected to compete in the finals on December 15. The finalists each presented their models and the judges tested them all privately. I was elated that my model won first prize, with the next two going to teams from Ho Chi Minh city.
Reporter: With its ability to recognize human emotions, your model can be applied in many different fields. What was your overall objective when you first designed it?
Mr. Tran Bao Toan: More and more complex models are being developed to assist in community work and healthcare applications. By designing emotion recognition models, I want to create alarms to help reduce traffic accidents. For example, by installing a device that recognizes the emotional condition of drivers, fatigue, drowsiness, drunkenness can be detected and an alarm sounded. In healthcare, the application can be used to detect medical symptoms by vocal analysis and help doctors draw up improved treatment regimens.
Reporter: How did you feel when you won first prize?
Mr. Tran Bao Toan: My feelings have not changed since! When my name was called out as the winner, I was very surprised, because the other models were all very impressive. This was the first contest I have taken part in and I gained a new insight and learned a lot from the other teams there. This experience will encourage me to continue working conscientiously in my current research. I also want to thank DTU for creating the hospitable and dynamic work environment that was conducive to entering the contest and winning.
(Media Center)