Should I use DSPs or FPGAs for implementation of speech and face recognition?

[harsha44]

hey guys, im planning to implement a speaker recognition or face recognition system on hardware. Do i use normal DSPs or FPGAs?
Also suggest some topic that integrates DSP and VLSI effectively for my final yr proj...Thanks

 

[ali_th]

need advice...

if you use dspic is very better.however this work is depend on your algoritem that you use for your prosesing.

 

[nikhr]

need advice...

hi..on which algorithm u want to work?accordingly u can select DSP or FPGA..

 

[mondunno]

need advice...

i think u're new comer to DSPs as well as FPGA.
if u have too much time to spend. go for fpga.
DSPs are easy to program and flexible in terms of built in hardware accelators.

first test the algorithm in matlab. then comes these 2

 

[harsha44]

Re: need advice...

ok.. im trying to implement a simple Digital synthesizer which i think might need some intensive calculations , thus justifying my need for an fpga..but as u said , a DSp is a simpler option.

And if uve heard of DDS, then im pretty sure u know abt phase dithering.I plan on implementing dithering with not just one noise signal, but more than one.This has the advantage of drastically reducing the size of the lookup table used in the algo.
So, any idea as to how i should go abt this?

Added after 3 minutes:

im sorry i said im implementing a face recognition system, thats another proj im working on parallelly...i need to learn some stuff and maybe ill have a better idea abt it..for now im focussing on the DDS.

Added after 2 minutes:

im sorry i said im implementing a face recognition system, thats another proj im working on parallelly...i need to learn some stuff and maybe ill have a better idea abt it..for now im focussing on the DDS.