INTERNATIONAL JOURNAL OF SCIENTIFIC DEVELOPMENT AND RESEARCH International Peer Reviewed & Refereed Journals, Open Access Journal ISSN Approved Journal No: 2455-2631 | Impact factor: 8.15 | ESTD Year: 2016
open access , Peer-reviewed, and Refereed Journals, Impact factor 8.15
Automatic speech emotion recognition has grown in popularity because it allows for natural human-computer connection. One way to recognize emotion is voice. Speech, however, also includes silence that cannot be related to emotion. The elimination of silence and/or ignoring silence while paying greater attention to the segment of speech is two ways to improve performance. This Paper propose a combination of silence elimination and a care model in this paper to enhance the performance of speech emotion. An improved CNN model is presented here which consists of combination of convolution 1d layers and generalized to form a 9 layer architecture of CNN (convolutional neural network), model accuracy has been checked with respect to emotion classes such as considering 5 emotions considered as angry, calm, fearful, happy, sad for male as well as female, likewise included use of classes such as positive, negative, neutral to achieve optimum accuracy. The results show that silence cancellation and attention model combinations are better than just the noise cancellation model or just the attention model. In the realm of human-computer interaction, speech emotion recognition is a critical and difficult job. Various models and feature sets for training the system have been proposed in previous work. Using input signals of various lengths, a novel speech-emotion detection system based on Convolutional Neural Networks (CNN) is presented in this research. With the use of a powerful GPU, a model is created and fed with unprocessed speech from a specified dataset for training, classification, and testing purposes. Finally, it achieves a convincing accuracy of 89.00%, which is far higher than any other comparable job on this dataset. This work will have an impact on the creation of social and conversational robots that can convey all the subtleties of human emotion.In terms of accuracy of the model the results are comparatively improved as compared to previous models using same dataset.
"Convolution neural network based Speech Emotion Recognition", International Journal of Science & Engineering Development Research (www.ijsdr.org), ISSN:2455-2631, Vol.9, Issue 1, page no.491 - 497, January-2024, Available :http://www.ijsdr.org/papers/IJSDR2401072.pdf
Downloads:
000338721
Publication Details:
Published Paper ID: IJSDR2401072
Registration ID:209874
Published In: Volume 9 Issue 1, January-2024
DOI (Digital Object Identifier):
Page No: 491 - 497
Publisher: IJSDR | www.ijsdr.org
ISSN Number: 2455-2631
Facebook Twitter Instagram LinkedIn