Fast signal transform with Banded matrix method
If y0u ever have had experiences for image, video, audio processing, you could know that the Fourier transform is used all the time. The question is even on a modern computer, it is still difficult to process the huge size of image, audio, video with Fourier transform.
MIT mathematics professor Gilbert Strang just made a breakthrough of this. In his paper titled “Fast transforms: Banded matrices with banded inverses”, he invented a new fast transform method called Banded matrices with banded inverses. This method is potentially much faster than Fourier transform for some applications, such as the special format of images, video, and audio, wireless signals.
Why not try it, if you are interested in developing super fast wireless communication, image, video, audio, text analysis software/hardware.
Original image (left) belongs to Flickr creative commons