Emotion Recognition in Women’s Voices and Its Application to Improve Mood Through Light Color Control

This paper presents a project focused on recognizing emotions in women’s voices and its application to improve mood through light color control. Using a dataset of 51 audio recordings of women saying “Good morning” with different emotions, a program was developed in MATLAB to recognize specific emotions in real-time. Tests were conducted with different age groups to evaluate the effectiveness of the program in identifying emotions. Additionally, a simulation of color change in an LED bulb was carried out and compared with another bulb controlled via Wi-Fi. While the first approach showed promising results, limitations were identified in Wi-Fi communication with the second bulb. The results indicate variable effectiveness in emotion recognition, with higher accuracy for sadness and anger compared to happiness. Despite current limitations due to data and resource availability, the project demonstrates the potential of technology to enhance emotional well-being through practical applications.