Speech Emotion Recognition Based on Speaker and Gender Dependences in Cross-Linguistic

Speech emotion recognition (SER) has received a great deal of attention. Because it is used in human-computer interaction. Emotional Computation F. This research explores the complex relationship between the domains of speech emotion recognition and speaker gender dependency within and across language contexts. DWT and SWT are used for preprocessing to increase the discriminative power of speech signals. These wavelet transforms can more effectively express the emotional information contained in speech signals. after pre-processing The corresponding frequency gain coefficient (MFCC) is extracted to capture important spectral characteristics of the speech signal. Multilevel Perceptron (MLP) and SVM models are used in the classification step. The performance accuracy of the running model is verified. The focus is on understanding speakers and sexual dependencies. Results indicate that the MLP model consistently outperforms the SVM model's accuracy by 74%, demonstrating 88% greater accuracy in within- and across-language situations. This study's pre-processing techniques Feature extraction method and classification models in the context of speaker and gender-dependent speech emotion recognition contribute valuable insights into this role.