Speech Enhancement Based on Deep AutoEncoder for Remote Arabic Speech Recognition

Abstract

Remote applications that deal with speech need the speech signal to be compressed. First, speech coding transforms the continuous waveform into a numerical form. Then, the digitized signal is compressed with or without loss of information. This transformation affects the original waveform and degrades performances for further recognition of the speech signal. Meanwhile, the transmission is another source of speech degradation. To restore the original "clean" speech, speech enhancement (SE) is widely used, and deep learning algorithms are state-of-the-art, nowadays. In this paper, the target application is a remote Arabic speech recognition system, and the aim of using SE is to improve the accuracy of the speech recognizer. For that purpose, a Deep Auto Encoder (DAE) is used. The effect of the DAE-based SE is studied through different configurations, and the performances are evaluated through accuracy. The results showed an improvement of about 3.17 between the accuracy prior to the SE and that computed with the enhanced speech.

Deep Learning Methods for Arabic Autoencoder Speech Recognition System for Electro-Larynx Device

“…After the denoising step with the wavelet technique, the next step is feature extraction. MFCC is one of the successful techniques for feature extraction [11,14,15]. Experiments revealed that MFCC is a commonly utilized technique, especially for a noisy dataset like the collected speech dataset produced by an EL device.…”

“…Model's name Accuracy Word error rate Jaber and Abdulbaqi [28] Autoencoder (CNN) 93% -Eljawad et al [7] Fuzzy neural network 94.5% -Dendani et al [11] Autoencoder (MLP) 65.72% -Alsayadi et al [14] Autoencoder (LSTM) 71.58 28.42% Alsayadi et al [15] CNN-LSTM -13.52% Proposed model Autoencoder (GRU) 95.31% 4.69%…”

“…Tey reported 65.0% accuracy for ASR in stressful conditions. In another research study by Dendani et al [11] an autoencoder model was proposed for enhancing Arabic language recognition performance. Tey used the proposed model to restore the original clean speech from noisy datasets.…”

Enabling Two-Way Communication of Deaf Using Saudi Sign Language

“…Long short-term memory (LSTM) and gated recurrent unit (GRU) are used for Arabic speech recognition [35]. Deep auto-encoder was used in [36], for remote Arabic speech recognition system. The authors used isolated words Arabic speech database for their experiments, where the database contained only a recording of 20 words.…”