【Python】keras 画像データセットの読み込み方 【DeepLearning】

多クラス分類　csvのラベルファイルがある場合

#ラべルの読み込み
LABELS = "../input/train.csv"
train_df = pd.read_csv(LABELS)
top = sorted(list(train_df['label_id'].value_counts().head(16).index))
train_df = train_df[train_df['label_id'].isin(top)]

from keras.preprocessing.image import img_to_array
from keras.preprocessing.image import load_img
from sklearn.model_selection import train_test_split

#データセットの成型
SEED = 1234
TRAIN_FOLDER = "../input/train/train/"
TEST_FOLDER = "../input/test/test/"
DIM = 299
train_df['image_path'] = train_df.apply( lambda x: ( TRAIN_FOLDER + x["file_name"] ), axis=1 )
train_data = np.array([ img_to_array(load_img(img, target_size=(DIM, DIM))) for img in train_df['image_path'].values.tolist()]).astype('float32')
train_labels = train_df['label_id']

#データ分割
x_train, x_validation, y_train, y_validation = train_test_split(train_data, train_labels, test_size=0.2, stratify=np.array(train_labels), random_state=SEED)
#ラベルをカテゴリ変数に変換
y_train = pd.get_dummies(y_train.reset_index(drop=True), columns=top_breeds).as_matrix()
y_validation = pd.get_dummies(y_validation.reset_index(drop=True), columns=top_breeds).as_matrix()

カテゴリがフォルダ毎に別れている場合

#globを使いデータを読み込み
z = glob.glob("../input/train/*/*.png")
ori_label = []
ori_imgs = []
for fn in z:
    if fn[-3:] != 'png':
        continue
    ori_label.append(fn.split('/')[-2])
    new_img = Image.open(fn)
    ori_imgs.append(ImageOps.fit(new_img, (48, 48), Image.ANTIALIAS).convert('RGB'))

#画像の配列化、正則化、カテゴリラベルの数値化
imgs = np.array([np.array(im) for im in ori_imgs])
imgs = imgs.reshape(imgs.shape[0], 48, 48, 3) / 255
lb = LabelBinarizer().fit(ori_label)
label = lb.transform(ori_label)
#データの分割
trainX, validX, trainY, validY = train_test_split(imgs, label, test_size=0.05, random_state=42)