def collect_video_frames(config):
cap = cv2.VideoCapture(config['input_files'][0])
base_path = config['output_dir']
counter = 0
images = []
files = []
while 1:
ret, frame = cap.read()
if ret == 0:
break
if counter % config['sampling_freq'] == 0:
image_name = f'{base_path}/frame_{counter}.png'
cv2.imwrite(image_name, frame)
frame = align_face(image_name)
frame.save(image_name)
images.append(cv2.imread(image_name))
files.append(image_name)
counter += 1
config['input_files'] = files
config['output_files'] = files
cap.release()
out = cv2.VideoWriter(f'{base_path}/frames.avi',
cv2.VideoWriter_fourcc(*'MJPG'),
config['video_freq'], (1024, 1024))
for file in files:
out.write(cv2.imread(file))
out.release()
def get_aligned(img_paths):
xs = []
for img_path in img_paths:
img = Image.open(img_path)
img = img.convert('RGB') # if image is RGBA or Grayscale etc
img = np.array(img)
x, face_found = align_face(img)
if face_found:
xs.append(x)
x = np.concatenate(xs, axis=0)
return x
def get_aligned(img_paths):
xs = []
for img_path in img_paths:
img = Image.open(img_path)
img = img.convert('RGB') # Nếu ảnh ở dạng RGBA hoặc Grayscale ,....
img = np.array(img)
x, face_found = align_face(img)
if face_found:
xs.append(x)
x = np.concatenate(xs, axis=0)
return x
def get_aligned(img_paths):
xs = []
c = 0
for img_path in img_paths:
c += 1
print(c)
img = Image.open(path + '/' + img_path)
img = img.resize((width, height), Image.NEAREST)
#print(img.shape)
img = img.convert('RGB') # if image is RGBA or Grayscale etc
img = np.array(img)
x, face_found = align_face(img)
if face_found:
xs.append(x)
x = np.concatenate(xs, axis=0)
return x
def align_encode():
r = request
img = get(r, 'img')
# img = parse_img(img) if in jpg etc format
img = deserialise_img(img)
img, face_found = align_face(img)
if face_found:
img = np.reshape(img, [1, 256, 256, 3])
print(img.shape)
z = model.encode(img)
proj = model.project(z) # get projections. Not used
result = img, z
# jsonify(img=serialise_img(img), z=serialise_nparr(z))
return send_proj(result, proj)
else:
return jsonify(face_found=False)
def align_encode():
r = request
img = get(r, 'img')
print(img)
# img = parse_img(img) if in jpg etc format
img = deserialise_img(img)
img, face_found = align_face(img)
if face_found:
img = np.reshape(img, [1, 256, 256, 3])
print(img.shape)
z = model.encode(img)
proj = model.project(z) # get projections. Not used
result = img, z
# jsonify(img=serialise_img(img), z=serialise_nparr(z))
return send_proj(result, proj)
else:
return jsonify(face_found=False)
import time
import dlib # requires cmake
import numpy as np
from align_face import align_face
from Library.Spout import Spout
spout = Spout(silent = False, width = 1044, height = 1088)
spout.createReceiver('input')
spout.createSender('output')
# python projector.py --target=out/seed0002.png --project-in-wplus --save-video --num-steps=1000 --network=https://api.ngc.nvidia.com/v2/models/nvidia/research/stylegan3/versions/1/files/stylegan3-r-afhqv2-512x512.pkl
while True :
# check on close window
spout.check()
# receive data
data = spout.receive()
print(data.shape)
data = align_face(data)
print(data.shape)
spout.send(data)
from pathlib import Path
orig_img_path = Path('imgs')
aligned_imgs_path = Path('aligned_imgs')
if not aligned_imgs_path.exists():
aligned_imgs_path.mkdir()
import dlib
from align_face import align_face
# Align all of our images using a landmark detection model!
all_imgs = list(orig_img_path.iterdir())
for img in all_imgs:
print(f'aligning {img}')
try:
align_face(str(img)).save(aligned_imgs_path / ('aligned_' + img.name))
except:
print('no face')
# aligned_img_set = list(aligned_imgs_path.iterdir())
# aligned_img_set.sort()
# aligned_img_set = [Image.open(x) for x in aligned_img_set]
# orig_img_set = list(orig_img_path.iterdir())
# orig_img_set.sort()
# orig_img_set = [Image.open(x) for x in orig_img_set]
def align_images(config):
for img_path in config['input_files']:
new_img = align_face(img_path)
new_img.save(img_path)