cnn_bounding_box.py 文件源码

def train(img_shape):
    # Model
    model = Sequential()

    model.add(
        Convolution2D(32, 3, 3, input_shape=img_shape, activation='relu', W_constraint=maxnorm(3), dim_ordering='tf'))
    model.add(Dropout(0.2))

    model.add(Convolution2D(32, 3, 3, activation='relu', W_constraint=maxnorm(3), dim_ordering='tf'))
    model.add(MaxPooling2D())

    model.add(Convolution2D(32, 3, 3, activation='relu', W_constraint=maxnorm(3), dim_ordering='tf'))
    model.add(MaxPooling2D())

    model.add(Convolution2D(32, 3, 3, activation='relu', W_constraint=maxnorm(3), dim_ordering='tf'))
    model.add(MaxPooling2D())

    model.add(Flatten())
    model.add(Dense(512, activation='relu', W_constraint=maxnorm(3)))
    model.add(Dropout(0.5))
    model.add(Dense(8))

    model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])
    for features, labels in feature_labels_generator():
        model.fit(features, labels, nb_epoch=1)
    # TODO: Get generator to
    # samples_per_epoch = 100
    # model.fit_generator(feature_labels_generator(), samples_per_epoch, nb_epoch=10)

    return model