Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

DSP, Classification & More (Proper Math Forum?)

  1. Jan 26, 2012 #1
    I always have an active imagination. Things that most people say are inconceivable I think of and wonder how to accomplish it. Luckily often before I hear it's "inconceivable." So I've been tinkering with this idea in my head. And one of my brother's ex-coworkers had done some work with this! So it is possible... just not sure if I posted in the right math forum.

    What this coworker did was develop classification algorithms for sonar systems at NUWC in Rhode Island. He drastically improved the performance by forming classification groups instead of having some brute force type of comparison. The trick was how he did it that I'm trying to understand. My brother talked about how the input signal had to be transformed to preserve the "energy" or convert the "energy" to a state that could be compared to the known signals.

    So my crazy idea is something similar with neural nets. It could be something similar as to what this guy did... using it on video signals instead to classify objects, etc. What I don't get is where to start even comprehending this. The idea is one thing... starting to implement it is another. What kind of math is involved? What should I be reading?

    I guess I have my work cut out for me because this dude was a Ph.D.

  2. jcsd
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook

Can you offer guidance or do you also need help?