# coding: utf-8

# In[1]:

get_ipython().magic('pylab inline')

import os

import numpy as np

import pandas as pd

from scipy.misc import imread

from sklearn.metrics import accuracy_score

import tensorflow as tf

import keras

# In[2]:

# To stop potential randomness

seed = 128

rng = np.random.RandomState(seed)

# In[3]:

import glob

from PIL import Image

import os.path

import re

from resizeimage import resizeimage

# In[4]:

data_dir = os.path.abspath('images_35_35/')

# In[5]:

os.path.exists(data_dir)

# In[6]:

train = pd.read_csv('train.csv',header = 0, sep = ',')

test = pd.read_csv('test.csv',header = 0, sep = ',')

sample_submission = pd.read_csv('sample_submission.csv', header = 0, sep = ",")

train.head()

# In[7]:

train.orientation.unique()

# In[8]:

train.shape

# In[9]:

img_name = rng.choice(train['id'])

img = imread('images_35_35/%d.jpg' % img_name, flatten=True)

pylab.imshow(img, cmap = 'gray')

pylab.axis('off')

pylab.show()

# In[10]:

img

# In[11]:

## Reading Train & Test

# In[12]:

temp = []

for img_name in train['id']:

img = imread('images_35_35/%d.jpg' % img_name, flatten=True)

temp.append(img)

train_x = np.stack(temp)

# In[13]:

train_x = train_x.reshape(9980, 35, 35, 1)

train_x = train_x.astype('float32')

train_x /= 255

# In[14]:

temp = []

for img_name in test['id']:

img = imread('images_35_35/%d.jpg' % img_name, flatten=True)

temp.append(img)

test_x = np.stack(temp)

# In[15]:

test_x = test_x.reshape(4124, 35, 35, 1)

test_x = test_x.astype('float32')

test_x /= 255

# In[16]:

train_y = train['orientation']

# In[17]:

train_y.head()

# In[18]:

num_classes = 4

# In[19]:

train_y = train_y -1

# In[20]:

train_y = keras.utils.to_categorical(train_y, num_classes)

# In[21]:

train_y

# In[22]:

## Splitting Train set into Train and Validation

# In[23]:

split_size = int(train_x.shape[0]*0.8)

# In[24]:

train_xx, val_x = train_x[:split_size], train_x[split_size:]

# In[25]:

train_yy, val_y = train_y[:split_size], train_y[split_size:]

# In[26]:

from keras.models import Sequential

from keras.optimizers import RMSprop, Adam

from keras.layers import Dense, Dropout, Activation, Flatten

from keras.layers.normalization import BatchNormalization

from keras.utils import np_utils

from keras.layers import Conv2D, MaxPooling2D, ZeroPadding2D, GlobalAveragePooling2D

from keras.layers.advanced_activations import LeakyReLU

from keras.preprocessing.image import ImageDataGenerator

# In[27]:

from keras.applications.vgg19 import VGG19

# In[194]:

# 1. Convolution

# 2. Activation

# 3. Pooling

# 4. Fully connected layer

model = Sequential()

model.add(Conv2D(32, (3, 3), input_shape=(35,35,1)))

model.add(BatchNormalization(axis=-1))

model.add(Activation('relu'))

model.add(Conv2D(64, (3, 3)))

model.add(BatchNormalization(axis=-1))

model.add(Activation('relu'))

model.add(MaxPooling2D(pool_size=(2,2)))

model.add(Conv2D(128,(3, 3)))

model.add(BatchNormalization(axis=-1))

model.add(Activation('relu'))

model.add(Conv2D(256, (3, 3)))

model.add(BatchNormalization(axis=-1))

model.add(Activation('relu'))

model.add(MaxPooling2D(pool_size=(2,2)))

model.add(Flatten())

# Fully connected layer

model.add(Dense(512))

model.add(BatchNormalization())

model.add(Activation('relu'))

model.add(Dropout(0.2))

model.add(Dense(4))

model.add(Activation('softmax'))

# In[195]:

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

# In[33]:

gen = ImageDataGenerator(rotation_range=8, width_shift_range=0.08, shear_range=0.3,

height_shift_range=0.08, zoom_range=0.08)

test_gen = ImageDataGenerator()

# In[36]:

train_generator = gen.flow(train_xx, train_yy)

test_generator = test_gen.flow(val_x, val_y)

# In[37]:

train_xx.shape

# In[38]:

val_x.shape

# In[203]:

model.fit_generator(train_generator, steps_per_epoch=7984//4, epochs=50, validation_data=test_generator, validation_steps=1996//4)

# In[204]:

score = model.evaluate(val_x, val_y, verbose=0)

# In[205]:

score

# In[181]:

test_x.shape

# In[182]:

pred = model.predict_proba(test_x)

# In[183]:

submit_df = pd.DataFrame(pred, columns=[1, 2, 3, 4])

# In[184]:

submit_df.head()

# In[185]:

submit_df.shape

# In[186]:

submit_df.to_csv('cnn_sub_15.csv')