model.py

import numpy as np
import os

import skimage.io as io
import skimage.transform as trans
import numpy as np
from keras.losses import binary_crossentropy
from keras.utils import multi_gpu_model
from utils import *
from keras import losses

from keras.models import *
from keras.layers import *
from keras.optimizers import *
from keras.utils import plot_model
from keras.callbacks import ModelCheckpoint, LearningRateScheduler

from keras import backend as K

import segmentation_models as sm
from segmodel import *

loss_dict = {
    "bceja": sm.losses.bce_jaccard_loss,
    "ja": sm.losses.jaccard_loss,
    "focal": sm.losses.binary_focal_loss,
    "bce":sm.losses.binary_crossentropy,
    "focalja": sm.losses.binary_focal_jaccard_loss,
    "focaldice":sm.losses.binary_focal_dice_loss,
    "dice": sm.losses.dice_loss,
    "bcedice": sm.losses.bce_dice_loss,
    "l1": losses.mean_absolute_percentage_error,
    "l2": losses.mean_squared_error
}

def unet(pretrained_weights=None, input_size=(256, 256, 3), lr=1E-3, multi_gpu=False, loss="l1"):
    # 所有等号左侧其实不是层而是张量吗...
    # 是的！ 因为这里使用了keras的函数式API。每一个层都是可以调用的...而在左边返回输出的张量
    def resblock(layer_input, filters, f_size=3):
        r = Conv2D(filters, kernel_size=f_size, strides=1, padding='same')(layer_input)
        r = LeakyReLU(alpha=0.2)(r)
        r = BatchNormalization()(r)
        r = Conv2D(filters, kernel_size=f_size, strides=1, padding='same')(r)
        r = LeakyReLU(alpha=0.2)(r)
        r = Add()([r, layer_input])
        return BatchNormalization()(r)

    def resblockn(n, layer_input, filters, f_size=3):
        x = layer_input
        for k in range(n):
            x = resblock(x, filters, f_size)
        return x

    inputs = Input(input_size)
    conv1 = Conv2D(64, 3, activation='relu', padding='same',
                   kernel_initializer='he_normal')(inputs)
    conv1 = Conv2D(64, 3, activation='relu', padding='same',
                   kernel_initializer='he_normal')(conv1)
    conv1 = BatchNormalization()(conv1)
    pool1 = MaxPooling2D(pool_size=(2, 2))(conv1)

    conv2 = Conv2D(128, 3, activation='relu', padding='same',
                   kernel_initializer='he_normal')(pool1)
    conv2 = Conv2D(128, 3, activation='relu', padding='same',
                   kernel_initializer='he_normal')(conv2)

    conv2 = BatchNormalization()(conv2)
    pool2 = MaxPooling2D(pool_size=(2, 2))(conv2)
    conv3 = Conv2D(256, 3, activation='relu', padding='same',
                   kernel_initializer='he_normal')(pool2)
    conv3 = Conv2D(256, 3, activation='relu', padding='same',
                   kernel_initializer='he_normal')(conv3)
    conv3 = BatchNormalization()(conv3)
    pool3 = MaxPooling2D(pool_size=(2, 2))(conv3)

    conv4 = Conv2D(512, 3, activation='relu', padding='same',
                   kernel_initializer='he_normal')(pool3)
    conv4 = Conv2D(512, 3, activation='relu', padding='same',
                   kernel_initializer='he_normal')(conv4)
    conv4 = BatchNormalization()(conv4)

    drop4 = Dropout(0.5)(conv4)
    pool4 = MaxPooling2D(pool_size=(2, 2))(drop4)

    conv5 = Conv2D(1024, 3, activation='relu', padding='same',
                   kernel_initializer='he_normal')(pool4)
    conv5 = Conv2D(1024, 3, activation='relu', padding='same',
                   kernel_initializer='he_normal')(conv5)
    conv5 = BatchNormalization()(conv5)

    drop5 = Dropout(0.5)(conv5)

    #res5 = resblockn(9, drop5, 1024)


    up6 = Conv2D(512, 2, activation='relu', padding='same',
                 kernel_initializer='he_normal')(UpSampling2D(size=(2, 2))(drop5))

    merge6 = concatenate([drop4, up6], axis=3)
    conv6 = Conv2D(512, 3, activation='relu', padding='same',
                   kernel_initializer='he_normal')(merge6)
    conv6 = Conv2D(512, 3, activation='relu', padding='same',
                   kernel_initializer='he_normal')(conv6)
    conv6 = BatchNormalization()(conv6)

    up7 = Conv2D(256, 2, activation='relu', padding='same',
                 kernel_initializer='he_normal')(UpSampling2D(size=(2, 2))(conv6))
    merge7 = concatenate([conv3, up7], axis=3)
    conv7 = Conv2D(256, 3, activation='relu', padding='same',
                   kernel_initializer='he_normal')(merge7)
    conv7 = Conv2D(256, 3, activation='relu', padding='same',
                   kernel_initializer='he_normal')(conv7)
    conv7 = BatchNormalization()(conv7)

    up8 = Conv2D(128, 2, activation='relu', padding='same',
                 kernel_initializer='he_normal')(UpSampling2D(size=(2, 2))(conv7))
    merge8 = concatenate([conv2, up8], axis=3)
    conv8 = Conv2D(128, 3, activation='relu', padding='same',
                   kernel_initializer='he_normal')(merge8)
    conv8 = Conv2D(128, 3, activation='relu', padding='same',
                   kernel_initializer='he_normal')(conv8)
    conv8 = BatchNormalization()(conv8)
    up9 = Conv2D(64, 2, activation='relu', padding='same',
                 kernel_initializer='he_normal')(UpSampling2D(size=(2, 2))(conv8))
    merge9 = concatenate([conv1, up9], axis=3)
    conv9 = Conv2D(64, 3, activation='relu', padding='same',
                   kernel_initializer='he_normal')(merge9)
    conv9 = Conv2D(64, 3, activation='relu', padding='same',
                   kernel_initializer='he_normal')(conv9)
    conv9 = Conv2D(2, 3, activation='relu', padding='same',
                   kernel_initializer='he_normal')(conv9)
    conv9 = BatchNormalization()(conv9)

    conv10 = Conv2D(1, 1, activation='sigmoid')(conv9)

    model = Model(inputs=inputs, outputs=conv10)
    if multi_gpu:
        strategy = tf.distribute.MirroredStrategy()
        with strategy.scope():
            model = multi_gpu_model(model, gpus=2)
            model.compile(optimizer=Adam(lr=lr),
                          loss=sm.losses.binary_focal_jaccard_loss, metrics=[sm.metrics.iou_score, 'accuracy'])
    else:
        model.compile(optimizer=Adam(lr=lr),
                      loss=loss_dict[loss], metrics=[sm.metrics.iou_score, 'accuracy'])

    # model.summary()
    if pretrained_weights:
        model.load_weights(pretrained_weights)

    return model


def smunet(loss="focal", pretrained_weights=None):
    model = sm.Unet(backbone_name = 'densenet121',
                    input_shape=(None, None, 3),
                    classes=1,
                    activation='sigmoid',
                    weights=None,
                    encoder_weights='imagenet',
                    encoder_freeze=False,
                    encoder_features='default',
                    decoder_block_type='upsampling',
                    decoder_filters=(256, 128, 64, 32, 16),
                    decoder_use_batchnorm=True)
    model.compile(optimizer='adam',
                loss=loss_dict[loss], metrics=[sm.metrics.iou_score, 'accuracy'])
    if pretrained_weights:
        model.load_weights(pretrained_weights)
    return model
model_dict = {
    "unet": unet,
    "unetxx": unetxx,
    "unet++": unetxx,
    "denseunet": denseunet
}