Selecting Sports Players Using Intelligent Techniques

Abstract

Speech is one of the most natural forms of expression. As technology advances, the interaction between humans and machines is becoming more sophisticated. Research has shown that emotions can be detected not only through facial expressions but also through sounds. Like facial features, auditory characteristics such as pitch and intensity can provide valuable information about a person\\'s emotional state. Studies suggest that it is possible to build intelligent systems that recognize individuals\\' emotional states through their voices, distinguishing between anger, sadness, happiness, fear, normal state, and others. These systems can be used in various applications such as driver safety systems and customer service systems. In this study, we aim to apply emotion recognition systems in the sports field. The developed system called CHOOSE_TEST that tests the emotional state of CHOOSE players before they play a tournament to determine their psychological stability, Negative feelings such as sadness and anger will be referred to as an ABNORMAL psychological state, and positive feelings such as feelings of joy and natural will be referred to as NORMAL psychological state. The system detects the player\\'s psychological state through sound and its features. We built two models for this system: DL_CHOOSE model based on CNN and ML_CHOOSE model based on pre-trained (ResNet50) and transfer learning concepts. The system has an accuracy rate of 97% using CNN and 87% using ResNet50 for discrimination purposes.