My project is to recognise the Arabic /english speech in a noisy environment using deep learning and hidden Markov models:
You are required to complete an application on Matlab that can train the database using deep learning, doing language model and lexical model for Arabic , adding functions to increase different types of noise at different levels as the picture shows. Then we notice the results.
The database that we will add to the test, whether through the micro or through the disk must be must be filtered and the noise is completely removed, we test it is clean and test it if we want with adding all types of noise.
for remove of noise in speech load use VAD method ,[login to view URL], and for filtration use 3 filtre or 4 ..
Therefore, a series of functions must be completed, which is accompanied by comments to clarify the code and how to work with it. This is done within five days if possible.