![]() Three hundred and sixty seven sentences are considered as phonetically rich and balanced, which are used for training Arabic Automatic Speech Recognition (ASR) systems. The speech corpus contains a total of 415 sentences recorded by 40 (20 male and 20 female) Arabic native speakers from 11 different Arab countries representing three major regions (Levant, Gulf, and Africa). This paper describes the preparation, recording, analyzing, and evaluation of a new speech corpus for Modern Standard Arabic (MSA). Keywords : Speech Recognition, Hidden Markov Models, EM Algorithm, Tagalog Alphabets. ![]() On the other hand, using the verification data, the recognition accuracy was 85.5%. A 99.0% and 92.0% recognition was achieved for female and male training data when (K, N) = (32, 6) and (K, N) = (36, 7) respectively. ![]() Five female and five male speakers were trained, varying the number of Vector Quantization (VQ) codebooks K (ranging from 20 to 40) and HMM states N (ranging from 2 to 10). Each window was feature extracted using the 12th order Linear Predictive Coding. Each overlapping frames has a window length of 400 samples (25ms) and windowing period is 120 samples (7.5ms). Speech was sampled at 16 kHz and was divided into frames tapered by Hamming window. ![]() The performance of the model was examined at training and recognition stages. An isolated word Tagalog alphabet recognizer is developed using the Hidden Markov Model (HMM). ABSTRACT Speech recognition is the process of transmuting human speech into text via a computer. ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |