python - Keras 中 NPY 格式的 4 通道图像的图像分类

Question

我正在使用卷积神经网络构建一个包含 3000 张图像的 5 类图像分类器。我的输入数据存储为具有形状 (1080,1080,4) 的 .npy 文件格式，这是通过 RGB 图像和边缘特征（4 通道）获得的。当尝试使用类 Datagenerator 训练模型时，我得到了训练准确度（acc train = 97%）和验证准确度（acc val = 51%）之间的巨大差异。您可以在下面找到我的代码摘录。谁能帮我解决这个问题，我该如何提高验证的准确性。

有什么方法可以使用 ImageDataGenerator 提供的数据增强功能？

非常感谢，任何帮助表示赞赏。

class DataGenerator(keras.utils.Sequence):
    'Generates data for Keras'
    def __init__(self, list_IDs, labels, batch_size=16, dim=(1080,1080), n_channels=4,
                 n_classes=5, shuffle=True):
        'Initialization'
        self.dim = dim
        self.batch_size = batch_size
        self.labels = labels
        self.list_IDs = list_IDs
        self.n_channels = n_channels
        self.n_classes = n_classes
        self.shuffle = shuffle
        self.on_epoch_end()
        
    def __len__(self):
        'Denotes the number of batches per epoch'
        return int(np.floor(len(self.list_IDs) / self.batch_size))

    def __getitem__(self, index):
        'Generate one batch of data'
        # Generate indexes of the batch
        indexes = self.indexes[index*self.batch_size:(index+1)*self.batch_size]

        # Find list of IDs
        list_IDs_temp = [self.list_IDs[k] for k in indexes]

        # Generate data
        X, y = self.__data_generation(list_IDs_temp)

        return X, y

    def on_epoch_end(self):
        'Updates indexes after each epoch'
        self.indexes = np.arange(len(self.list_IDs))
        if self.shuffle == True:
            np.random.shuffle(self.indexes)
            

    def __data_generation(self, list_IDs_temp):
        'Generates data containing batch_size samples' # X : (n_samples, *dim, n_channels)
        # Initialization
        X = np.empty((self.batch_size, *self.dim, self.n_channels))
        y = np.empty((self.batch_size), dtype=int)

        # Generate data
        for i, ID in enumerate(list_IDs_temp):
            # Store sample
            X[i,] = np.load('data' + ID + '.npy')
            # Store class
            y[i] = self.labels[ID]
            
        return X, keras.utils.to_categorical(y, num_classes=self.n_classes)

还有模型：

model = Sequential()
# Note the input shape is the desired size of the image img_width, img_height with 3 bytes color
# 1____The first convolution
model.add(Convolution2D(16, (3, 3), input_shape=(1080,1080,4)))
model.add(Activation('relu'))
model.add(MaxPooling2D(pool_size=(2,2))) # ,dim_ordering="th")

# 2____The second convolution
model.add(Convolution2D(32, (3, 3)))
model.add(Activation('relu'))
model.add(MaxPooling2D(pool_size=(2, 2)))

# 3____The third convolution
model.add(Convolution2D(64, (3, 3)))
model.add(Activation('relu'))
model.add(MaxPooling2D(pool_size=(2,2)))

# 4____The fourth convolution
model.add(Convolution2D(64, (3, 3)))
model.add(Activation('relu'))
model.add(MaxPooling2D(pool_size=(2,2)))

# 5____The fifth convolution
model.add(Convolution2D(64, (3, 3)))
model.add(Activation('relu'))
model.add(MaxPooling2D(pool_size=(2,2)))

# F____Flatten the results to feed into a dense layer
model.add(Flatten())    # this converts our 3D feature maps to 1D feature vectors

# D____128 neuron in the fully-connected layer
model.add(Dense(128))
model.add(Activation('relu'))

# D____Dropout
model.add(Dropout(0.5))

# D____5 output neurons for 5 classes with the softmax activation
model.add(Dense(5))
model.add(Activation('softmax'))

model.summary()

# model compile
opt = Adam(lr=0.001, decay=1e-6)
model.compile(loss='categorical_crossentropy',
              optimizer=opt,
              metrics=['acc'])

标签： pythontensorflowkeras