Sigproc function
Bilateral is a noise reduction function.
See the file test_pwc.py for an example.
bilateral
-
sigproc.bilateral(y, soft, beta, width, display=1, stoptol=1.0, maxiter=50)
Performs PWC denoising of the input signal using hard or soft kernel
bilateral filtering.
- Parameters:
- y: numpy array
- Original signal to denoise of length N.
- soft: int
- Set this to 1 to use the soft Gaussian kernel, else uses
the hard kernel.
- beta: int
- Kernel parameter. If soft Gaussian kernel, then this is the
precision parameter. If hard kernel, this is the kernel
support.
- width: int
- Spatial kernel width W.
- display: int [default 1]
- Set to 0 to turn off progress display, 1 to turn
on. If not specifed, defaults to progress display on.
- stoptol: float [default 1**-3]
- Precision of estimate as determined by square
magnitude of the change in the solution. If not specified,
defaults to 1e-3.
- maxiter: int [default 50]
- Maximum number of iterations. If not specified,
defaults to 50.
- Results: numpy array
- Denoised output signal.
- Reference:
(c) Max Little, 2011. If you use this code for your research, please cite:
M.A. Little, Nick S. Jones (2011)
“Generalized Methods and Solvers for Noise Removal from Piecewise
Constant Signals: Part I and II”
Proceedings of the Royal Society A (in press).
See http://www.maxlittle.net/