PCNN

Python implementation of the Pulse Coupled Neural Network (PCNN) alongside multiple variations:

• Classical PCNN
• Feature Linking Model (FLM)
• Intersecting Cortical Model (ICM)
• Multi Linking Model (MLM)
• Spiking Cortical Model (SCM)
• Sigmoidal Linking Model (SLM)

Install:

pip install pulse_coupled_nn

Usage example:

import numpy as np
import matplotlib.pyplot as plt

from pulse_coupled_nn import FLM
from pulse_coupled_nn import ICM
from pulse_coupled_nn import ClassicalPCNN
from pulse_coupled_nn import SCM
from pulse_coupled_nn import SLM


def run_image_segm(gamma=1, beta=2, v_theta=400, kernel_size=3, kernel='gaussian'):

    image = np.array(
        [[230, 230, 230, 230, 115, 115, 115, 115],
        [230, 230, 230, 230, 115, 115, 115, 115],
        [230, 230, 205, 205, 103, 103, 115, 115],
        [230, 230, 205, 205, 103, 103, 115, 115],
        [230, 230, 205, 205, 103, 103, 115, 115],
        [230, 230, 230, 230, 115, 115, 115, 115],
        [230, 230, 230, 230, 115, 115, 115, 115]]
    )

    model = ClassicalPCNN(image.shape, kernel, kernel_size=kernel_size)
    segm_image = model.segment_image(image, gamma=gamma, beta=beta, v_theta=v_theta, kernel_type='gaussian')

    plt.imshow(image)
    plt.colorbar()
    plt.show()

    plt.imshow(segm_image)
    plt.colorbar()
    plt.show()


run_image_segm()