Convolutional network used to quantify and identify bacteria colony patches.

mycnn.py

from keras.models import Sequential
from keras.layers.convolutional import Conv2D, MaxPooling2D, ZeroPadding2D
from keras.constraints import maxnorm
from keras.layers import Flatten, Dense, Dropout, Activation
from keras import backend as K
from keras.layers.normalization import BatchNormalization
from keras.regularizers import l2

def mycnn(img_shape=(32, 32, 1), n_classes=2, l2_reg=0.0, weights=None):
    K.set_image_data_format("channels_last")
    # Initialize model
    model = Sequential()
    # Layer 1
    model.add(Conv2D(96, (3, 3), input_shape=img_shape, padding='same', kernel_regularizer=l2(l2_reg)))
    model.add(BatchNormalization())
    model.add(Activation('relu'))

    # Layer 2
    model.add(Conv2D(128, (3, 3), padding='same'))
    model.add(BatchNormalization())
    model.add(Activation('relu'))

    # Layer 3
    model.add(ZeroPadding2D((1, 1)))
    model.add(Conv2D(256, (3, 3), padding='same'))
    model.add(BatchNormalization())
    model.add(Activation('relu'))
    model.add(MaxPooling2D(pool_size=(2, 2)))

    # Layer 4
    model.add(ZeroPadding2D((1, 1)))
    model.add(Conv2D(256, (3, 3), padding='same'))
    model.add(BatchNormalization())
    model.add(Activation('relu'))
    model.add(MaxPooling2D(pool_size=(2, 2)))

    # Layer 5
    model.add(ZeroPadding2D((1, 1)))
    model.add(Conv2D(256, (3, 3), padding='same'))
    model.add(BatchNormalization())
    model.add(Activation('relu'))
    model.add(MaxPooling2D(pool_size=(2, 2)))

    # Layer 6
    model.add(Flatten())
    model.add(Dense(256))  # 3072
    model.add(BatchNormalization())
    model.add(Activation('relu'))
    model.add(Dropout(0.5))

    # Layer 7
    model.add(Dense(128))  # 4096 # 1024
    model.add(BatchNormalization())
    model.add(Activation('relu'))
    model.add(Dropout(0.5))

    # Layer 8
    model.add(Dense(n_classes))
    model.add(BatchNormalization())
    model.add(Activation('softmax'))

    if weights is not None:
        model.load_weights(weights)

    return model