Cat Hipsterizer

이탤릭 볼드 이탤릭볼드

Workflow stages

Question or problem definition.
Acquire training and testing data.
Wrangle, prepare, cleanse the data.
Analyze, identify patterns, and explore the data.
Model, predict and solve the problem.
Visualize, report, and present the problem solving steps and final solution.
Supply or submit the results.

기본적으로 설치되어 있어야하는 패키지는 아래 코드 를 사용한다.

import random, sys
import dlib, cv2, os
import pandas as pd
import numpy as np

# Train 


# visualization
import seaborn as sns
import matplotlib.pyplot as plt
%matplotlib inline

# machine learning
from sklearn.linear_model import LogisticRegression
from sklearn.svm import SVC, LinearSVC
from sklearn.ensemble import RandomForestClassifier
from sklearn.neighbors import KNeighborsClassifier
from sklearn.naive_bayes import GaussianNB
from sklearn.linear_model import Perceptron
from sklearn.linear_model import SGDClassifier
from sklearn.tree import DecisionTreeClassifier

data 가져오기

dirname = 'CAT_00'
base_path = 'C:/Users/k1014/PycharmProjects/cat_hipsterizer/images/%s' % dirname
file_list = sorted(os.listdir(base_path))

data를 찍어보면 다음과 같이 나온다

cat file을 불러서 landmark 값을 pandas로 9 * 2 형태로 바꿔줌

for f in file_list:
  if '.cat' not in f:
    continue

  # read landmarks
  pd_frame = pd.read_csv(os.path.join(base_path, f), sep=' ', header=None)
  landmarks = (pd_frame.as_matrix()[0][1:-1]).reshape((-1, 2)).astype(np.int)

  # load image
  img_filename, ext = os.path.splitext(f)

  img = cv2.imread(os.path.join(base_path, img_filename))

  # visualize 이미지안에 landmark를 붉은색(0, 0, 225)으로 표시
  for l in landmarks:
    cv2.circle(img, center=tuple(l), radius=1, color=(0, 0, 255), thickness=2)

  cv2.imshow('img', img)
  if cv2.waitKey(0) == ord('q'):
    break

전처리

img_size = 224
dirname = 'CAT_00'
base_path = 'C:/Users/k1014/PycharmProjects/cat_hipsterizer/images/%s' % dirname
file_list = sorted(os.listdir(base_path))

# random.shuffle(file_list)

# 최종적으로 dataset 을 저장할 변수
dataset = {
  'imgs': [],
  'lmks': [],
  'bbs': []
}

# 이미지를 resize함
def resize_img(im):
  old_size = im.shape[:2] # old_size is in (height, width) format
  ratio = float(img_size) / max(old_size)
  new_size = tuple([int(x*ratio) for x in old_size])
  # new_size should be in (width, height) format
  im = cv2.resize(im, (new_size[1], new_size[0]))
  delta_w = img_size - new_size[1]
  delta_h = img_size - new_size[0]
  top, bottom = delta_h // 2, delta_h - (delta_h // 2)
  left, right = delta_w // 2, delta_w - (delta_w // 2)
  new_im = cv2.copyMakeBorder(im, top, bottom, left, right, cv2.BORDER_CONSTANT,
      value=[0, 0, 0]) # 비어있는 부분은 검은색으로 칠함
  return new_im, ratio, top, left

# cat 파일만 불러오도록 필터링
for f in file_list:
  if '.cat' not in f:
    continue

  # read landmarks
  pd_frame = pd.read_csv(os.path.join(base_path, f), sep=' ', header=None)
  landmarks = (pd_frame.as_matrix()[0][1:-1]).reshape((-1, 2)) # 9 * 2 행렬로 변환

  # load image
  img_filename, ext = os.path.splitext(f)

  img = cv2.imread(os.path.join(base_path, img_filename))

  # resize image and relocate landmarks
  img, ratio, top, left = resize_img(img)
  landmarks = ((landmarks * ratio) + np.array([left, top])).astype(np.int) # 변한 landmark를 재 계산
  bb = np.array([np.min(landmarks, axis=0), np.max(landmarks, axis=0)]) # 얼굴의 영역. landmark의 최소점, 최대점을 구함

  dataset['imgs'].append(img)
  dataset['lmks'].append(landmarks.flatten())
  dataset['bbs'].append(bb.flatten())

  # for l in landmarks:
  #   cv2.circle(img, center=tuple(l), radius=1, color=(255, 255, 255), thickness=2)

  # cv2.imshow('img', img)
  # if cv2.waitKey(0) == ord('q'):
  #   break

np.save('dataset/%s.npy' % dirname, np.array(dataset)) # 마지막 전처리된 데이터 저장

CAT_00.npy … 파일들이 저장됨

모델 생성

# modeling
'''
inputs = Input(shape=(img_size, img_size, 3))

mobilenetv2_model = mobilenetv2.MobileNetV2(input_shape=(img_size, img_size, 3), alpha=1.0, depth_multiplier=1, include_top=False, weights='imagenet', input_tensor=inputs, pooling='max')

net = Dense(128, activation='relu')(mobilenetv2_model.layers[-1].output)
net = Dense(64, activation='relu')(net)
net = Dense(output_size, activation='linear')(net)

model = Model(inputs=inputs, outputs=net)
'''

# modeling tf 2.x 로 변환 mobilenetv2
# include_top = False 최종 Output Layer들을 포함시키지 않고 모델 로드
# pooling -> average or max
inputs = tf.keras.layers.Input(shape=(img_size, img_size, 3))
model = tf.keras.Sequential([
    tf.keras.applications.MobileNetV2(input_shape=(img_size, img_size, 3), alpha=1.0, include_top=False, weights='imagenet',input_tensor=inputs, pooling='max'),
    tf.keras.layers.Dense(128, activation='relu'),
    tf.keras.layers.Dense(64, activation='relu'),
    tf.keras.layers.Dense(output_size, activation='linear')
])

model.summary()

컴파일

# training
model.compile(optimizer=tf.keras.optimizers.Adam(), loss='mse')

model.fit(x_train, y_train, epochs=50, batch_size=32, shuffle=True,
  validation_data=(x_test, y_test), verbose=1,
  callbacks=[
    tf.keras.TensorBoard(log_dir='logs/%s' % (start_time)),
    tf.keras.ModelCheckpoint('./models/%s.h5' % (start_time), monitor='val_loss', verbose=1, save_best_only=True, mode='auto'),
    tf.keras.ReduceLROnPlateau(monitor='val_loss', factor=0.2, patience=5, verbose=1, mode='auto')
  ]
)

Cat Hipsterizer

Cat Hipsterizer

Workflow stages

data 가져오기

전처리

모델 생성

컴파일

Bumso KIM

Comments