Human emotions play a significant role in communication, and speech is probably the best of all natural media for expressing emotions. Speech Emotion Recognition is a technology that recognizes and classes emotions from a speaker’s voice. SER systems are interested in the automation of emotion detection and classification from speech signals. A number of applications of this field can be identified such as human–computer interaction, healthcare, customer service, and security, among others. It still remains a challenge to recognize emotions from speech objectively and accurately because it is a very complex and variable thing between and among individuals, languages, and environments. Advances in artificial intelligence and machine learning made recognitions technology take huge steps forward. The central problem in Speech Emotion Recognition is the design of an automatic system that could reasonably classify a speaker’s emotional state based solely on raw speech data. In most of the approaches developed to overfitting, poor generalization to unseen data, and Deficiencies in feature extraction lead to suboptimal. Additionally, variability of speech signals due to noise, speaker individuality, and cultural differences makes SER system building pretty complicated.