METHOD AND SYSTEM FOR PERFORMING SPEECH EMOTION RECOGNITION USING SPECTRUM, AND STORAGE MEDIUM

摘要

The present application relates to the technical field of artificial intelligence, and provides a method and system for performing speech emotion recognition using a spectrum, and a storage medium. The method comprises: pre-processing speech; performing spectrum calculation on the pre-processed speech and obtaining spectral values; normalizing the spectral values; and using a deep neural network to perform emotion recognition on the normalized spectral values, and obtaining an emotion recognition result. The present application can be used to resolve prior-art issues such as a low speech emotion recognition rate, time delay, and poor robustness to unidentified environmental noise.