Anime2Sketchでアニメ画像から線画への変換を実装する
/
Anime2Sketchでアニメ画像から線画への変換を実装します．
全国630店舗以上！もみほぐし・足つぼ・ハンドリフレ・クイックヘッドのリラクゼーション店【りらくる】
# Google Driveと連携
Google ColabとGoogle Driveを連携させて，gitからMukosame/Anime2Sketch (opens new window)をダウンロードします．
# Google ColabとGoogle Driveを連携
from google.colab import drive
drive.mount('/content/drive')
# ディレクトリの移動
%cd /content/drive/MyDrive/
# gitのダウンロード
!git clone https://github.com/Mukosame/Anime2Sketch.git
!ls
# 作業ディレクトリへの移動
%cd /content/drive/MyDrive/Anime2Sketch
!ls
# モジュールのインストール
下記のコードでモジュールのインストールします．
!pip install -r requirements.txt
# 学習済みモデルのダウンロード
下記のリンクから学習済みモデルのnetG.pthをダウンロードして，weightsフォルダーに入れてください．
GoogleDrive (opens new window)
%%bash
# 確認
cd weights
ls
# netG.pth, Put the weights here
# コードの書き換え
エラーが発生したため，下記のコードの変更します．
try:
from torchvision.transforms import InterpolationMode
bic = InterpolationMode.BICUBIC
except ImportError:
bic = Image.BICUBIC
# try:
# from torchvision.transforms import InterpolationMode
# bic = InterpolationMode.BICUBIC
# except ImportError:
bic = Image.BICUBIC
%%writefile data.py
import os
from PIL import Image
import torchvision.transforms as transforms
# try:
# from torchvision.transforms import InterpolationMode
# bic = InterpolationMode.BICUBIC
# except ImportError:
bic = Image.BICUBIC
import numpy as np
import torch
IMG_EXTENSIONS = ['.jpg', '.JPG', '.jpeg', '.JPEG', '.png', '.PNG', '.ppm', '.PPM', '.bmp', '.BMP']
def is_image_file(filename):
"""if a given filename is a valid image
Parameters:
filename (str) -- image filename
"""
return any(filename.endswith(extension) for extension in IMG_EXTENSIONS)
def get_image_list(path):
"""read the paths of valid images from the given directory path
Parameters:
path (str) -- input directory path
"""
assert os.path.isdir(path), '{:s} is not a valid directory'.format(path)
images = []
for dirpath, _, fnames in sorted(os.walk(path)):
for fname in sorted(fnames):
if is_image_file(fname):
img_path = os.path.join(dirpath, fname)
images.append(img_path)
assert images, '{:s} has no valid image file'.format(path)
return images
def get_transform(load_size=0, grayscale=False, method=bic, convert=True):
transform_list = []
if grayscale:
transform_list.append(transforms.Grayscale(1))
if load_size > 0:
osize = [load_size, load_size]
transform_list.append(transforms.Resize(osize, method))
if convert:
transform_list += [transforms.ToTensor()]
if grayscale:
transform_list += [transforms.Normalize((0.5,), (0.5,))]
else:
transform_list += [transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5))]
return transforms.Compose(transform_list)
def read_img_path(path, load_size):
"""read tensors from a given image path
Parameters:
path (str) -- input image path
load_size(int) -- the input size. If <= 0, don't resize
"""
img = Image.open(path).convert('RGB')
aus_resize = None
if load_size > 0:
aus_resize = img.size
transform = get_transform(load_size=load_size)
image = transform(img)
return image.unsqueeze(0), aus_resize
def tensor_to_img(input_image, imtype=np.uint8):
""""Converts a Tensor array into a numpy image array.
Parameters:
input_image (tensor) -- the input image tensor array
imtype (type) -- the desired type of the converted numpy array
"""
if not isinstance(input_image, np.ndarray):
if isinstance(input_image, torch.Tensor): # get the data from a variable
image_tensor = input_image.data
else:
return input_image
image_numpy = image_tensor[0].cpu().float().numpy() # convert it into a numpy array
if image_numpy.shape[0] == 1: # grayscale to RGB
image_numpy = np.tile(image_numpy, (3, 1, 1))
image_numpy = (np.transpose(image_numpy, (1, 2, 0)) + 1) / 2.0 * 255.0 # post-processing: tranpose and scaling
else: # if it is a numpy array, do nothing
image_numpy = input_image
return image_numpy.astype(imtype)
def save_image(image_numpy, image_path, output_resize=None):
"""Save a numpy image to the disk
Parameters:
image_numpy (numpy array) -- input numpy array
image_path (str) -- the path of the image
output_resize(None or tuple) -- the output size. If None, don't resize
"""
image_pil = Image.fromarray(image_numpy)
if output_resize:
print(output_resize)
print(bic)
image_pil = image_pil.resize(size=output_resize, resample=bic)
image_pil.save(image_path)
# 画像での変換を実行
下記のコードで画像での変換を実行します．
# テスト実行
!python3 test.py --dataroot test_samples/madoka.jpg --load_size 512 --output_dir results/
下記のコードでJupyter上で画像を確認します．
from IPython.display import Image,display_jpeg
display_jpeg(Image('test_samples/madoka.jpg'))
from IPython.display import Image,display_jpeg
display_jpeg(Image('results/madoka.jpg'))
# 動画での変換を実行
下記のコードで動画での変換するpythonスクリプトを作成します．
%%writefile test_video.py
import os
import cv2
import numpy as np
import torch
import torchvision.transforms as transforms
from data import get_image_list, get_transform
from model import create_model
from data import read_img_path, tensor_to_img, save_image
import argparse
from PIL import Image
bic = Image.BICUBIC
if __name__ == '__main__':
parser = argparse.ArgumentParser(description='Anime-to-sketch test options.')
parser.add_argument('--dataroot','-i', default='test_samples/', type=str)
parser.add_argument('--load_size','-s', default=512, type=int)
parser.add_argument('--output_video','-o', default='results/sample.mp4', type=str)
parser.add_argument('--gpu_ids', '-g', default=[], help="gpu ids: e.g. 0 0,1,2 0,2.")
opt = parser.parse_args()
# create model
gpu_list = ','.join(str(x) for x in opt.gpu_ids)
os.environ['CUDA_VISIBLE_DEVICES'] = gpu_list
device = torch.device('cuda' if len(opt.gpu_ids)>0 else 'cpu')
model = create_model().to(device) # create a model given opt.model and other options
model.eval()
# get input data
if os.path.isfile(opt.dataroot):
# load video
cap = cv2.VideoCapture(opt.dataroot)
# video info
codec = cv2.VideoWriter_fourcc(*'XVID')
cap_fps = int(cap.get(cv2.CAP_PROP_FPS))
cap_width,cap_height= int(cap.get(cv2.CAP_PROP_FRAME_WIDTH)), int(cap.get(cv2.CAP_PROP_FRAME_HEIGHT))
# output video
output_video = cv2.VideoWriter(opt.output_video, codec, cap_fps, (cap_width,cap_height))
else:
raise Exception("{} is not a valid directory or video file.".format(opt.dataroot))
# save outputs
# save_dir = opt.output_video
# os.makedirs(save_dir, exist_ok=True)
# for test_path in test_list:
while True:
ret, frame = cap.read()
if frame is None:
print('Completed')
break
# basename = os.path.basename(test_path)
# aus_path = os.path.join(save_dir, basename)
# video frame image
# img, aus_resize = read_img_path(test_path, opt.load_size)
aus_resize = None
frame = Image.fromarray(frame) # cv2 -> pillow
if opt.load_size > 0:
aus_resize = frame.size
transform = get_transform(load_size=opt.load_size)
image = transform(frame)
img = image.unsqueeze(0)
# model convert
aus_tensor = model(img.to(device))
aus_img = tensor_to_img(aus_tensor)
# resize image
# save_image(aus_img, aus_path, aus_resize)
image_pil = Image.fromarray(aus_img)
image_pil = image_pil.resize(size=aus_resize, resample=bic)
# output video AND cv2 -> pillow
image_cv2 = np.array(image_pil, dtype=np.uint8)
output_video.write(cv2.cvtColor(image_cv2, cv2.COLOR_BGR2RGB))
cap.release()
output_video.release()
下記のコードを動画での変換を実施します．
# 変換実行
!python3 test_video.py --dataroot ./video/vinland.mp4 --load_size 512 --output_video ./video/out_vinland.mp4
# まとめ
Anime2Sketchでアニメ画像から線画への変換を実装しました．
# 参考サイト
Mukosame/Anime2Sketch (opens new window)
