[머신러닝]합성곱 신경망 모델 구현

손글씨 원, 사각형, 삼각형 분류

라이브 러리들

import numpy as np
from keras.models import Sequential
from keras.layers import Dense
from keras.layers import Flatten
from keras.layers.convolutional import Conv2D
from keras.layers.convolutional import MaxPooling2D
from keras.preprocessing.image import ImageDataGenerator
import matplotlib.pyplot as plt

데이터 전처리

np.random.seed(3)

train_datagen = ImageDataGenerator(rescale=1./255)

train_generator = train_datagen.flow_from_directory(
    '/content/drive/MyDrive/handwriting_shape/train',
    target_size=(24,24),
    batch_size = 3,
    class_mode='categorical'
)

test_datagen = ImageDataGenerator(rescale=1./255)

test_generator = test_datagen.flow_from_directory(
    '/content/drive/MyDrive/handwriting_shape/test',
    target_size=(24,24),
    batch_size = 3,
    class_mode='categorical'
)

모델구성

model = Sequential()
model.add(Conv2D(32, kernel_size=(3,3), activation = 'relu', input_shape=(24,24,3)))
model.add(Conv2D(64,(3,3),activation = 'relu'))
#model.add(MaxPooling2D(pool_size=(2,2)))
model.add(Flatten())
model.add(Dense(128, activation='relu'))
model.add(Dense(3,activation='softmax'))

모델 컴파일

model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])

모델 학습

model.fit_generator(train_generator,steps_per_epoch=15,epochs=50,validation_data=test_generator,validation_steps=5)

Epoch 49/50
15/15 [==============================] - 1s 43ms/step - loss: 5.2955e-06 - accuracy: 1.0000 - val_loss: 0.2440 - val_accuracy: 0.9333
Epoch 50/50
15/15 [==============================] - 1s 36ms/step - loss: 4.9405e-06 - accuracy: 1.0000 - val_loss: 0.2473 - val_accuracy: 0.9333

모델 평가

print("--Evaluate--")
score= model.evaluate_generator(test_generator, steps=5)
print(score)

--Evaluate--
[0.24732321500778198, 0.9333333373069763]

모델 예측

print("--predict")
output = model.predict_generator(test_generator, steps=5)
np.set_printoptions(formatter={'float': lambda x: "{0:0.3f}".format(x)})
print(test_generator.class_indices)
for i in range( len(output)):
    print(np.argmax(output[i]))

--predict
{'circle': 0, 'rectangle': 1, 'triangle': 2}
0
0
2
0
2
2
1
1
2
1
1
0
0
2
1

---

데이터들