Keras 2 : examples : 画像分類のための CutMix データ増強 (翻訳/解説)
翻訳 : (株)クラスキャット セールスインフォメーション
作成日時 : 11/16/2021 (keras 2.7.0)
* 本ページは、Keras の以下のドキュメントを翻訳した上で適宜、補足説明したものです：
- Code examples : Computer Vision : CutMix data augmentation for image classification (Author: Sayan Nath)
* サンプルコードの動作確認はしておりますが、必要な場合には適宜、追加改変しています。
* ご自由にリンクを張って頂いてかまいませんが、sales-info@classcat.com までご一報いただけると嬉しいです。
クラスキャット 人工知能 研究開発支援サービス ★ 無料 Web セミナー開催中 ★
◆ クラスキャットは人工知能・テレワークに関する各種サービスを提供しております。お気軽にご相談ください :
- 人工知能研究開発支援
- 人工知能研修サービス(経営者層向けオンサイト研修)
- テクニカルコンサルティングサービス
- 実証実験(プロトタイプ構築)
- アプリケーションへの実装
- 人工知能研修サービス
- PoC(概念実証)を失敗させないための支援
- テレワーク & オンライン授業を支援
◆ 人工知能とビジネスをテーマに WEB セミナーを定期的に開催しています。スケジュール。
- お住まいの地域に関係なく Web ブラウザからご参加頂けます。事前登録 が必要ですのでご注意ください。
- ウェビナー運用には弊社製品「ClassCat® Webinar」を利用しています。
◆ お問合せ : 本件に関するお問い合わせ先は下記までお願いいたします。
- 株式会社クラスキャット セールス・マーケティング本部 セールス・インフォメーション
- E-Mail：sales-info@classcat.com ; WebSite: www.classcat.com ; Facebook
Keras 2 : examples : 画像分類のための CutMix データ増強
Description: CIFAR-10 上の画像分類のための CutMix によるデータ増強。
イントロダクション
CutMix はデータ増強テクニックで、領域 (= regional) dropout ストラテジーに存在する情報損失と非効率性の問題に対処します。ピクセルを削除してそれらを黒やグレーのピクセルか Gaussian ノイズで埋める代わりに、削除された領域を別の画像からのパッチで置き換えます、その一方で正解ラベルは結合された画像のピクセル数に比例してミックスされます。CutMix は CutMix: Regularization Strategy to Train Strong Classifiers with Localizable Features (Yun et al., 2019) で提案されました。
次の式で実装されています :
ここで M は二値マスクで、2 つのランダムにドローされた画像からの cutout (切り出し) と fill-in 領域を示し、そして λ (in [0, 1]) は Beta(α, α) 分布 からドローされます。
バウンディングボックスの座標は以下です :
これは画像の場合の cutout と fill-in 領域を示します。バウンディングボックスのサンプリングは以下で表されます :
ここで rx, ry は上界を持つ一様分布からランダムにドローされます。
セットアップ
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import tensorflow as tf
from tensorflow import keras
np.random.seed(42)
tf.random.set_seed(42)
CIFAR-10 データセットをロードする
この例では、CIFAR-10 画像分類データセットを使用します。
(x_train, y_train), (x_test, y_test) = tf.keras.datasets.cifar10.load_data()
y_train = tf.keras.utils.to_categorical(y_train, num_classes=10)
y_test = tf.keras.utils.to_categorical(y_test, num_classes=10)
print(x_train.shape)
print(y_train.shape)
print(x_test.shape)
print(y_test.shape)
class_names = [
"Airplane",
"Automobile",
"Bird",
"Cat",
"Deer",
"Dog",
"Frog",
"Horse",
"Ship",
"Truck",
]
(50000, 32, 32, 3) (50000, 10) (10000, 32, 32, 3) (10000, 10)
ハイパーパラメータの定義
AUTO = tf.data.AUTOTUNE
BATCH_SIZE = 32
IMG_SIZE = 32
画像前処理関数の定義
def preprocess_image(image, label):
image = tf.image.resize(image, (IMG_SIZE, IMG_SIZE))
image = tf.image.convert_image_dtype(image, tf.float32) / 255.0
return image, label
データを TensorFlow Dataset オブジェクトに変換する
train_ds_one = (
tf.data.Dataset.from_tensor_slices((x_train, y_train))
.shuffle(1024)
.map(preprocess_image, num_parallel_calls=AUTO)
)
train_ds_two = (
tf.data.Dataset.from_tensor_slices((x_train, y_train))
.shuffle(1024)
.map(preprocess_image, num_parallel_calls=AUTO)
)
train_ds_simple = tf.data.Dataset.from_tensor_slices((x_train, y_train))
test_ds = tf.data.Dataset.from_tensor_slices((x_test, y_test))
train_ds_simple = (
train_ds_simple.map(preprocess_image, num_parallel_calls=AUTO)
.batch(BATCH_SIZE)
.prefetch(AUTO)
)
# Combine two shuffled datasets from the same training data.
train_ds = tf.data.Dataset.zip((train_ds_one, train_ds_two))
test_ds = (
test_ds.map(preprocess_image, num_parallel_calls=AUTO)
.batch(BATCH_SIZE)
.prefetch(AUTO)
)
CutMix データ増強関数の定義
CutMix 関数は増強を実行するために 2 つの画像とラベルのペアを取ります。それは Beta 分布 から λ(l) をサンプリングして get_box 関数からバウンディング・ボックスを返します。そして 2 番目の画像 (image2) をクロップしてこの画像を同じ位置で最終的なパッドされた画像内でパッドします。
def sample_beta_distribution(size, concentration_0=0.2, concentration_1=0.2):
gamma_1_sample = tf.random.gamma(shape=[size], alpha=concentration_1)
gamma_2_sample = tf.random.gamma(shape=[size], alpha=concentration_0)
return gamma_1_sample / (gamma_1_sample + gamma_2_sample)
@tf.function
def get_box(lambda_value):
cut_rat = tf.math.sqrt(1.0 - lambda_value)
cut_w = IMG_SIZE * cut_rat # rw
cut_w = tf.cast(cut_w, tf.int32)
cut_h = IMG_SIZE * cut_rat # rh
cut_h = tf.cast(cut_h, tf.int32)
cut_x = tf.random.uniform((1,), minval=0, maxval=IMG_SIZE, dtype=tf.int32) # rx
cut_y = tf.random.uniform((1,), minval=0, maxval=IMG_SIZE, dtype=tf.int32) # ry
boundaryx1 = tf.clip_by_value(cut_x[0] - cut_w // 2, 0, IMG_SIZE)
boundaryy1 = tf.clip_by_value(cut_y[0] - cut_h // 2, 0, IMG_SIZE)
bbx2 = tf.clip_by_value(cut_x[0] + cut_w // 2, 0, IMG_SIZE)
bby2 = tf.clip_by_value(cut_y[0] + cut_h // 2, 0, IMG_SIZE)
target_h = bby2 - boundaryy1
if target_h == 0:
target_h += 1
target_w = bbx2 - boundaryx1
if target_w == 0:
target_w += 1
return boundaryx1, boundaryy1, target_h, target_w
@tf.function
def cutmix(train_ds_one, train_ds_two):
(image1, label1), (image2, label2) = train_ds_one, train_ds_two
alpha = [0.25]
beta = [0.25]
# Get a sample from the Beta distribution
lambda_value = sample_beta_distribution(1, alpha, beta)
# Define Lambda
lambda_value = lambda_value[0][0]
# Get the bounding box offsets, heights and widths
boundaryx1, boundaryy1, target_h, target_w = get_box(lambda_value)
# Get a patch from the second image (`image2`)
crop2 = tf.image.crop_to_bounding_box(
image2, boundaryy1, boundaryx1, target_h, target_w
)
# Pad the `image2` patch (`crop2`) with the same offset
image2 = tf.image.pad_to_bounding_box(
crop2, boundaryy1, boundaryx1, IMG_SIZE, IMG_SIZE
)
# Get a patch from the first image (`image1`)
crop1 = tf.image.crop_to_bounding_box(
image1, boundaryy1, boundaryx1, target_h, target_w
)
# Pad the `image1` patch (`crop1`) with the same offset
img1 = tf.image.pad_to_bounding_box(
crop1, boundaryy1, boundaryx1, IMG_SIZE, IMG_SIZE
)
# Modify the first image by subtracting the patch from `image1`
# (before applying the `image2` patch)
image1 = image1 - img1
# Add the modified `image1` and `image2` together to get the CutMix image
image = image1 + image2
# Adjust Lambda in accordance to the pixel ration
lambda_value = 1 - (target_w * target_h) / (IMG_SIZE * IMG_SIZE)
lambda_value = tf.cast(lambda_value, tf.float32)
# Combine the labels of both images
label = lambda_value * label1 + (1 - lambda_value) * label2
return image, label
Note : 2 つの画像を単一の画像を作成するために組合せています。
CutMix 増強を適用後に新しいデータセットを可視化する
# Create the new dataset using our `cutmix` utility
train_ds_cmu = (
train_ds.shuffle(1024)
.map(cutmix, num_parallel_calls=AUTO)
.batch(BATCH_SIZE)
.prefetch(AUTO)
)
# Let's preview 9 samples from the dataset
image_batch, label_batch = next(iter(train_ds_cmu))
plt.figure(figsize=(10, 10))
for i in range(9):
ax = plt.subplot(3, 3, i + 1)
plt.title(class_names[np.argmax(label_batch[i])])
plt.imshow(image_batch[i])
plt.axis("off")
ResNet-20 モデルの定義
def resnet_layer(
inputs,
num_filters=16,
kernel_size=3,
strides=1,
activation="relu",
batch_normalization=True,
conv_first=True,
):
conv = keras.layers.Conv2D(
num_filters,
kernel_size=kernel_size,
strides=strides,
padding="same",
kernel_initializer="he_normal",
kernel_regularizer=keras.regularizers.l2(1e-4),
)
x = inputs
if conv_first:
x = conv(x)
if batch_normalization:
x = keras.layers.BatchNormalization()(x)
if activation is not None:
x = keras.layers.Activation(activation)(x)
else:
if batch_normalization:
x = keras.layers.BatchNormalization()(x)
if activation is not None:
x = keras.layers.Activation(activation)(x)
x = conv(x)
return x
def resnet_v20(input_shape, depth, num_classes=10):
if (depth - 2) % 6 != 0:
raise ValueError("depth should be 6n+2 (eg 20, 32, 44 in [a])")
# Start model definition.
num_filters = 16
num_res_blocks = int((depth - 2) / 6)
inputs = keras.layers.Input(shape=input_shape)
x = resnet_layer(inputs=inputs)
# Instantiate the stack of residual units
for stack in range(3):
for res_block in range(num_res_blocks):
strides = 1
if stack > 0 and res_block == 0: # first layer but not first stack
strides = 2 # downsample
y = resnet_layer(inputs=x, num_filters=num_filters, strides=strides)
y = resnet_layer(inputs=y, num_filters=num_filters, activation=None)
if stack > 0 and res_block == 0: # first layer but not first stack
# linear projection residual shortcut connection to match
# changed dims
x = resnet_layer(
inputs=x,
num_filters=num_filters,
kernel_size=1,
strides=strides,
activation=None,
batch_normalization=False,
)
x = keras.layers.add([x, y])
x = keras.layers.Activation("relu")(x)
num_filters *= 2
# Add classifier on top.
# v1 does not use BN after last shortcut connection-ReLU
x = keras.layers.AveragePooling2D(pool_size=8)(x)
y = keras.layers.Flatten()(x)
outputs = keras.layers.Dense(
num_classes, activation="softmax", kernel_initializer="he_normal"
)(y)
# Instantiate model.
model = keras.models.Model(inputs=inputs, outputs=outputs)
return model
def training_model():
return resnet_v20((32, 32, 3), 20)
initial_model = training_model()
initial_model.save_weights("initial_weights.h5")
CutMix で増強されたデータセットでモデルを訓練する
model = training_model()
model.load_weights("initial_weights.h5")
model.compile(loss="categorical_crossentropy", optimizer="adam", metrics=["accuracy"])
model.fit(train_ds_cmu, validation_data=test_ds, epochs=15)
test_loss, test_accuracy = model.evaluate(test_ds)
print("Test accuracy: {:.2f}%".format(test_accuracy * 100))
Epoch 1/15 1563/1563 [==============================] - 62s 24ms/step - loss: 1.9216 - accuracy: 0.4090 - val_loss: 1.9737 - val_accuracy: 0.4061 Epoch 2/15 1563/1563 [==============================] - 37s 24ms/step - loss: 1.6549 - accuracy: 0.5325 - val_loss: 1.5033 - val_accuracy: 0.5061 Epoch 3/15 1563/1563 [==============================] - 38s 24ms/step - loss: 1.5536 - accuracy: 0.5840 - val_loss: 1.2913 - val_accuracy: 0.6112 Epoch 4/15 1563/1563 [==============================] - 38s 24ms/step - loss: 1.4988 - accuracy: 0.6097 - val_loss: 1.0587 - val_accuracy: 0.7033 Epoch 5/15 1563/1563 [==============================] - 38s 24ms/step - loss: 1.4531 - accuracy: 0.6291 - val_loss: 1.0681 - val_accuracy: 0.6841 Epoch 6/15 1563/1563 [==============================] - 37s 24ms/step - loss: 1.4173 - accuracy: 0.6464 - val_loss: 1.0265 - val_accuracy: 0.7085 Epoch 7/15 1563/1563 [==============================] - 37s 24ms/step - loss: 1.3932 - accuracy: 0.6572 - val_loss: 0.9540 - val_accuracy: 0.7331 Epoch 8/15 1563/1563 [==============================] - 37s 24ms/step - loss: 1.3736 - accuracy: 0.6680 - val_loss: 0.9877 - val_accuracy: 0.7240 Epoch 9/15 1563/1563 [==============================] - 38s 24ms/step - loss: 1.3575 - accuracy: 0.6782 - val_loss: 0.8944 - val_accuracy: 0.7570 Epoch 10/15 1563/1563 [==============================] - 38s 24ms/step - loss: 1.3398 - accuracy: 0.6886 - val_loss: 0.8598 - val_accuracy: 0.7649 Epoch 11/15 1563/1563 [==============================] - 38s 24ms/step - loss: 1.3277 - accuracy: 0.6939 - val_loss: 0.9032 - val_accuracy: 0.7603 Epoch 12/15 1563/1563 [==============================] - 38s 24ms/step - loss: 1.3131 - accuracy: 0.6964 - val_loss: 0.7934 - val_accuracy: 0.7926 Epoch 13/15 1563/1563 [==============================] - 37s 24ms/step - loss: 1.3050 - accuracy: 0.7029 - val_loss: 0.8737 - val_accuracy: 0.7552 Epoch 14/15 1563/1563 [==============================] - 37s 24ms/step - loss: 1.2987 - accuracy: 0.7099 - val_loss: 0.8409 - val_accuracy: 0.7766 Epoch 15/15 1563/1563 [==============================] - 37s 24ms/step - loss: 1.2953 - accuracy: 0.7099 - val_loss: 0.7850 - val_accuracy: 0.8014 313/313 [==============================] - 3s 9ms/step - loss: 0.7850 - accuracy: 0.8014 Test accuracy: 80.14%
(訳注 : 実験結果)
Epoch 1/15 1563/1563 [==============================] - 52s 22ms/step - loss: 1.9147 - accuracy: 0.4117 - val_loss: 1.7649 - val_accuracy: 0.4317 Epoch 2/15 1563/1563 [==============================] - 34s 22ms/step - loss: 1.6647 - accuracy: 0.5343 - val_loss: 1.4942 - val_accuracy: 0.5248 Epoch 3/15 1563/1563 [==============================] - 34s 22ms/step - loss: 1.5595 - accuracy: 0.5816 - val_loss: 1.1121 - val_accuracy: 0.6726 Epoch 4/15 1563/1563 [==============================] - 33s 21ms/step - loss: 1.4971 - accuracy: 0.6103 - val_loss: 0.9899 - val_accuracy: 0.7235 Epoch 5/15 1563/1563 [==============================] - 33s 21ms/step - loss: 1.4462 - accuracy: 0.6354 - val_loss: 0.9944 - val_accuracy: 0.7145 Epoch 6/15 1563/1563 [==============================] - 33s 21ms/step - loss: 1.4165 - accuracy: 0.6475 - val_loss: 0.9900 - val_accuracy: 0.7167 Epoch 7/15 1563/1563 [==============================] - 34s 22ms/step - loss: 1.3854 - accuracy: 0.6646 - val_loss: 0.9780 - val_accuracy: 0.7147 Epoch 8/15 1563/1563 [==============================] - 33s 21ms/step - loss: 1.3651 - accuracy: 0.6727 - val_loss: 0.9889 - val_accuracy: 0.7269 Epoch 9/15 1563/1563 [==============================] - 33s 21ms/step - loss: 1.3548 - accuracy: 0.6776 - val_loss: 0.8532 - val_accuracy: 0.7676 Epoch 10/15 1563/1563 [==============================] - 32s 21ms/step - loss: 1.3376 - accuracy: 0.6869 - val_loss: 0.8521 - val_accuracy: 0.7835 Epoch 11/15 1563/1563 [==============================] - 33s 21ms/step - loss: 1.3270 - accuracy: 0.6923 - val_loss: 0.9160 - val_accuracy: 0.7428 Epoch 12/15 1563/1563 [==============================] - 33s 21ms/step - loss: 1.3117 - accuracy: 0.7015 - val_loss: 0.8464 - val_accuracy: 0.7819 Epoch 13/15 1563/1563 [==============================] - 33s 21ms/step - loss: 1.3075 - accuracy: 0.7036 - val_loss: 0.9161 - val_accuracy: 0.7486 Epoch 14/15 1563/1563 [==============================] - 34s 22ms/step - loss: 1.2993 - accuracy: 0.7057 - val_loss: 0.8246 - val_accuracy: 0.7829 Epoch 15/15 1563/1563 [==============================] - 35s 22ms/step - loss: 1.2878 - accuracy: 0.7125 - val_loss: 0.8477 - val_accuracy: 0.7713 313/313 [==============================] - 2s 6ms/step - loss: 0.8477 - accuracy: 0.7713 Test accuracy: 77.13%
元の非増強データセットを使用してモデルを訓練する
model = training_model()
model.load_weights("initial_weights.h5")
model.compile(loss="categorical_crossentropy", optimizer="adam", metrics=["accuracy"])
model.fit(train_ds_simple, validation_data=test_ds, epochs=15)
test_loss, test_accuracy = model.evaluate(test_ds)
print("Test accuracy: {:.2f}%".format(test_accuracy * 100))
Epoch 1/15 1563/1563 [==============================] - 38s 23ms/step - loss: 1.4864 - accuracy: 0.5173 - val_loss: 1.3694 - val_accuracy: 0.5708 Epoch 2/15 1563/1563 [==============================] - 36s 23ms/step - loss: 1.0682 - accuracy: 0.6779 - val_loss: 1.1424 - val_accuracy: 0.6686 Epoch 3/15 1563/1563 [==============================] - 36s 23ms/step - loss: 0.8955 - accuracy: 0.7449 - val_loss: 1.0555 - val_accuracy: 0.7007 Epoch 4/15 1563/1563 [==============================] - 36s 23ms/step - loss: 0.7890 - accuracy: 0.7878 - val_loss: 1.0575 - val_accuracy: 0.7079 Epoch 5/15 1563/1563 [==============================] - 36s 23ms/step - loss: 0.7107 - accuracy: 0.8175 - val_loss: 1.1395 - val_accuracy: 0.7062 Epoch 6/15 1563/1563 [==============================] - 36s 23ms/step - loss: 0.6524 - accuracy: 0.8397 - val_loss: 1.1716 - val_accuracy: 0.7042 Epoch 7/15 1563/1563 [==============================] - 36s 23ms/step - loss: 0.6098 - accuracy: 0.8594 - val_loss: 1.4120 - val_accuracy: 0.6786 Epoch 8/15 1563/1563 [==============================] - 36s 23ms/step - loss: 0.5715 - accuracy: 0.8765 - val_loss: 1.3159 - val_accuracy: 0.7011 Epoch 9/15 1563/1563 [==============================] - 36s 23ms/step - loss: 0.5477 - accuracy: 0.8872 - val_loss: 1.2873 - val_accuracy: 0.7182 Epoch 10/15 1563/1563 [==============================] - 36s 23ms/step - loss: 0.5233 - accuracy: 0.8988 - val_loss: 1.4118 - val_accuracy: 0.6964 Epoch 11/15 1563/1563 [==============================] - 36s 23ms/step - loss: 0.5165 - accuracy: 0.9045 - val_loss: 1.3741 - val_accuracy: 0.7230 Epoch 12/15 1563/1563 [==============================] - 36s 23ms/step - loss: 0.5008 - accuracy: 0.9124 - val_loss: 1.3984 - val_accuracy: 0.7181 Epoch 13/15 1563/1563 [==============================] - 36s 23ms/step - loss: 0.4896 - accuracy: 0.9190 - val_loss: 1.3642 - val_accuracy: 0.7209 Epoch 14/15 1563/1563 [==============================] - 36s 23ms/step - loss: 0.4845 - accuracy: 0.9231 - val_loss: 1.5469 - val_accuracy: 0.6992 Epoch 15/15 1563/1563 [==============================] - 36s 23ms/step - loss: 0.4749 - accuracy: 0.9294 - val_loss: 1.4034 - val_accuracy: 0.7362 313/313 [==============================] - 3s 9ms/step - loss: 1.4034 - accuracy: 0.7362 Test accuracy: 73.62%
Epoch 1/15 1563/1563 [==============================] - 27s 16ms/step - loss: 1.4801 - accuracy: 0.5209 - val_loss: 1.3486 - val_accuracy: 0.5796 Epoch 2/15 1563/1563 [==============================] - 24s 15ms/step - loss: 1.0666 - accuracy: 0.6785 - val_loss: 1.3312 - val_accuracy: 0.6086 Epoch 3/15 1563/1563 [==============================] - 25s 16ms/step - loss: 0.8947 - accuracy: 0.7449 - val_loss: 1.1774 - val_accuracy: 0.6657 Epoch 4/15 1563/1563 [==============================] - 24s 15ms/step - loss: 0.7868 - accuracy: 0.7884 - val_loss: 0.9961 - val_accuracy: 0.7232 Epoch 5/15 1563/1563 [==============================] - 25s 16ms/step - loss: 0.7119 - accuracy: 0.8181 - val_loss: 1.1473 - val_accuracy: 0.7092 Epoch 6/15 1563/1563 [==============================] - 24s 15ms/step - loss: 0.6525 - accuracy: 0.8432 - val_loss: 1.3620 - val_accuracy: 0.6622 Epoch 7/15 1563/1563 [==============================] - 24s 15ms/step - loss: 0.6069 - accuracy: 0.8607 - val_loss: 1.3751 - val_accuracy: 0.6663 Epoch 8/15 1563/1563 [==============================] - 25s 16ms/step - loss: 0.5704 - accuracy: 0.8785 - val_loss: 1.4153 - val_accuracy: 0.6654 Epoch 9/15 1563/1563 [==============================] - 24s 15ms/step - loss: 0.5424 - accuracy: 0.8896 - val_loss: 1.2651 - val_accuracy: 0.7131 Epoch 10/15 1563/1563 [==============================] - 24s 15ms/step - loss: 0.5210 - accuracy: 0.8990 - val_loss: 1.1913 - val_accuracy: 0.7343 Epoch 11/15 1563/1563 [==============================] - 24s 16ms/step - loss: 0.5108 - accuracy: 0.9069 - val_loss: 1.6186 - val_accuracy: 0.6890 Epoch 12/15 1563/1563 [==============================] - 24s 15ms/step - loss: 0.4935 - accuracy: 0.9166 - val_loss: 1.3162 - val_accuracy: 0.7338 Epoch 13/15 1563/1563 [==============================] - 24s 15ms/step - loss: 0.4853 - accuracy: 0.9218 - val_loss: 1.2976 - val_accuracy: 0.7425 Epoch 14/15 1563/1563 [==============================] - 24s 15ms/step - loss: 0.4845 - accuracy: 0.9248 - val_loss: 1.6508 - val_accuracy: 0.6919 Epoch 15/15 1563/1563 [==============================] - 25s 16ms/step - loss: 0.4829 - accuracy: 0.9280 - val_loss: 1.8357 - val_accuracy: 0.6959 313/313 [==============================] - 2s 6ms/step - loss: 1.8357 - accuracy: 0.6959 Test accuracy: 69.59%
ノート
このサンプルでは、モデルを 15 エポック訓練しました。私達の実験では、CutMix を使用したモデルは、増強を使用しないモデル (72.70%) と比較して、 CIFAR-10 データセット上でより良い精度を達成しました (実験では 80.36%)。CutMix 増強を使用してモデルを訓練するには少ない時間ですむことに気付くかもしれません。
元の論文 に従って CutMix により更に実験でけいます。
以上