Hi!
Iam new to speech recognition..........
We want to do a simple speech recognition (like 4 or 6 words) using BF533EZ kit
lite....
I have seen some of the implementations using ATmega32 processor on web,based on
them i want to do speech recognition.They used filters(Low pass,High Pass and
Bandpass) to filter the finite duration speech and then prepared fingerprints.By
comparing the Dictionary fingerprints with Taken samples they are recognizing
the samples.They used 8Khz sampling frequency. For each sample they applied 8 to
10 filters and accumulated the outputs,like this they have done for .5
seconds(word size)...........
I think there will be different implementations for ex. in DSP we will take a
frame of samples (ex 256) to process rather tan individual samples.....
Is there any other different implementation for doing speech recognition in real
time even simple and small project also not a problem....
