Taqi's Magic Words 1.0b
Magic Words is my most current Project. It involves the development of a user-friendly speech recognition engine that performs better than the stupid program I wasted a lot of money on. The project owes a lot to Microsoft Research as I didn’t know a thing about Speech recognition before I read their site. The program uses their speech recognition engine and as I just have an evaluation license, the engine is not included in the program.
Mission Statement:
Although there are a lot of speech recognition programs popping up, none performs like the HAL 2000 ideal we all have in mind when we open the box. For one, we cannot 'speak' to a dormant computer and wake it up. Computers are already too difficult to use and PC-Magazine agrees. These products are adding a lot of complexity to an already complex machine. But most importantly, the speech recognition engines are good, they are becoming more accurate, but I DIDN’T MAKE THEM! So now I'm making my own.
Help appreciated:
Making the program is a challenge I have accepted, but help/suggestions are always welcome. The areas I am experiencing problems in are:
Accuracy in SR
Product Usability -The UI design is actually what I've been struggling with
Piracy protection- I have come up with an indigenous plan of action but not everybody agrees with its practicability.
How do you select training scripts so that a user only has to train the computer in what he is more likely to say?
Back to homepage | Burhan College Webcenter | Cadet College Hasanabdal Page