16611022
The novelty of introducing a "Nevisa System" system as the first continuous independent continuous speech recognition system for the Persian language. Conventional and customized techniques for various module systems are combined. For each module, modification and parameter optimization are required. The parameters specified for each section of the system are found by separately evaluating the performance of that section with different parameter values.
This system is developed in the process of academic and industrial teamwork and is intended to be a product that can be exploited. Therefore, noisy environment issues and speaker variations should be addressed. Various resistance techniques are pursued and optimized for this purpose. We also customize and use statistical and grammatical language models for Persian languages. General n-gram statistics of Persian language are taken and incorporated for the first time. Our evaluation results and real environmental tests show that the performance system is quite satisfactory for use by regular users.
No comments:
Post a Comment