Enhanced Emotion Analysis of Voice Feedback using CNN

In today’s rapidly evolving business landscape, customer feedback plays a pivotal role in shaping product development, service improvement, and overall customer satisfaction. While text-based sentiment analysis has gained prominence, it often lacks the depth and variation required to fully understand customer emotions. Our proposal is a system that aims to revolutionize the way businesses interpret customer feedback by analyzing the emotions conveyed through voice recordings. The proposal leverages modern speech processing technologies, including speech-emotion recognition (SER) Mel- Spectrogram and Convolutional Neural Network (CNN), to convert spoken words into spectrogram. Subsequently, it employs advanced emotional intelligence algorithms to detect and analyze the underlying emotions within the spoken content. This comprehensive approach goes beyond mere sentiment analysis, allowing businesses to gain a deeper insight into their customers’ emotional states, such as happiness, frustration, anger, satisfaction, and more which further improves customer feedback.