· Must be based on C#.Net and SqlServer DB
· The system must be based on machine learning approach
· The system must be based on efficient speach recognition algorithms hidden markove model
· Work with the most very well known format of voice files
· User can identify and store the Arabic keywords in the DB, and the application allow user to store different pronounce by different peoples for the same keyword.
· The application access the folder of voice files and search in each voice file and detect which file include the searched keyword and how many time
· The application show the result as table includes files name and how many this keyword included in this file, and the parts of voice had this keywords.
· The report no less than 4 pages and must include the following sections:
o Overview of voice recognition algorithms
o Related work
o Our application include flowchart how it is work
o Experimental and results from the application