Constant Fraction DiscriminatorΒΆ

The CFD is an algorithm for doing peak detection, usually implemented in hardware. Here, we look at doing an implementation in Python.

import numpy as np
from scipy.ndimage.interpolation import shift

def cfd(signal, fraction, delay=1, threshold=0):
    """Constant Fraction Discriminator

    :param np.array signal: the input from the digitizer
    :param float fraction: the fraction, from 0 to 1
    :param float delay: the delay of the CFD
    :param float threshold: the threshold
    :return np.array edge_indices: the start and stop indices for each peak

    # Scaled and delayed signal
    scaled = signal * fraction
    delayed = shift(scaled, -delay, mode="nearest")
    cfd_signal = signal - delayed

    # Edge detection
    edges_bool = cfd_signal > threshold
    edges_bool_2 = edges_bool[1:] != edges_bool[:-1]
    edge_indices = np.squeeze(np.where(edges_bool_2 == True))

    return edge_indices

Last updated: 01/12/2018