Arabic Phonemes Recognition Using Convolutional Neural Network
| Autor: | Irwan Mazlin, Fariza Hanis Abdul Razak, Zan Azma Nasruddin, Wan Adilah Wan Adnan |
|---|---|
| Rok vydání: | 2019 |
| Předmět: |
0209 industrial biotechnology
Signal processing Computer science Speech recognition Feature extraction 02 engineering and technology Pronunciation Convolutional neural network Class (biology) ComputingMethodologies_PATTERNRECOGNITION 020901 industrial engineering & automation Cross entropy Computer Science::Sound Multilayer perceptron Classifier (linguistics) 0202 electrical engineering electronic engineering information engineering 020201 artificial intelligence & image processing |
| Zdroj: | Communications in Computer and Information Science ISBN: 9789811503986 SCDS |
| DOI: | 10.1007/978-981-15-0399-3_21 |
| Popis: | This paper focuses on a machine learning that learn the correct pronunciation Arabic phonemes. In this study, the researchers develop using convolutional neural network as feature extraction in order to enhance the performance of the model and Multi layer perceptron as the classifier to classify classes. Different parameters of CNN model are used in order to investigate the best parameter for the recognition purpose. The dataset have been recorded from experts using smartphone which consist of 880 recorded audios to train the model (210 for each class). The researchers have experimented the models to measure the accuracy and the cross entropy in the training process. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|