def fimage(x, y, file, type=1):
if file[:3] == '/sd':
utils.filecp(file, '/flash/fcache', blocksize=8192)
lcd.image(x, y, '/flash/fcache', 0, type)
os.remove('/flash/fcache')
else:
lcd.image(x, y, file, 0, type)
def main():
remote.synchronise()
files = parser.parse_file_list(remote.get_content())
while len(files) == 0:
print("No files found, retrying...")
remote.synchronise()
files = parser.parse_file_list(remote.get_content())
time.sleep(20)
while True:
for file in files:
if file.type == parser.FileType.IMAGE:
display.image(file)
elif file.type == parser.FileType.VIDEO:
display.video(file)
else:
print("Unknown file type of " + str(file.path))
def showPic(imgs, locs, win=0, name="Real"):
imgs = [(x + 0.5) / 2 * 255 for x in imgs]
imgs = [cv2.flip(x.transpose(1, 2, 0), 0) for x in imgs]
# imgs = [x.transpose(1, 2, 0) for x in imgs]
for i in range(4):
for y in range(16):
for x in range(16):
if locs[i][y][x] > 0.3:
cv2.rectangle(imgs[i], (x * 8, 127 - y * 8),
(x * 8 + 8, 119 - y * 8), (0, 0, 255), 1)
# cv2.rectangle(imgs[i], (x * 8, y * 8), (x * 8 + 8, y * 8 + 8), (0, 0, 255), 1)
half = len(imgs) // 2
row1 = np.concatenate(imgs[:half], 1)
row2 = np.concatenate(imgs[half:], 1)
content = np.concatenate((row1, row2), 0)
try:
display.image(content, win=win, title=name)
except Exception as e:
print(e)
import random
import time
import display
def generate_image():
X, Y = numpy.meshgrid(numpy.linspace(0, numpy.pi, 512),
numpy.linspace(0, 2, 512))
z = (numpy.sin(X) + numpy.cos(Y))**2 + 0.5
return z
i1 = generate_image()
i2 = generate_image()
display.image(i1, title='gradient')
# display.images([i2, i2, i2, i2], width=200, title='super fabio', labels=['a', 'b', 'c', 'd'])
data = []
for i in range(15):
data.append([i, random.random(), random.random() * 2])
win = display.plot(data, labels=['position', 'a', 'b'], title='progress')
for i in range(15, 25):
time.sleep(0.2)
data.append([i, random.random(), random.random() * 2])
display.plot(data, win=win)
out_folder_name = time_info.strftime("%Y-%m-%d") + '_' \
+ input_name[:-4] \
+ '_2D' + time_info.strftime("_%H%M")
if not os.path.exists('./Trained_models/' + out_folder_name):
os.mkdir('./Trained_models/' + out_folder_name)
copyfile('./train_g2d_periodic.py',
'./Trained_models/' + out_folder_name + '/code.txt')
# load images
input_texture = Image.open('./Images/' + input_name)
input_torch = Variable(prep(input_texture)).unsqueeze(0).cuda()
# display images
if disp:
img_disp = numpy.asarray(input_texture, dtype="int32")
display.image(img_disp, win='input', title='Input texture')
#define layers, loss functions, weights and compute optimization target
out_keys = ['r11', 'r21', 'r31', 'r41', 'r51']
loss_fns = [GramMSELoss()] * len(out_keys)
loss_fns = [loss_fn.cuda() for loss_fn in loss_fns]
# these are the weights settings recommended by Gatys
# to use with Gatys' normalization:
# w = [1e2/n**3 for n in [64,128,256,512,512]]
w = [1, 1, 1, 1, 1]
#compute optimization targets
targets = [GramMatrix()(f).detach() for f in vgg_net(input_torch, out_keys)]
# training parameters
batch_size = 10
#!/usr/bin/env python
import numpy
import random
import time
import display
def generate_image():
X, Y = numpy.meshgrid(numpy.linspace(0, numpy.pi, 512), numpy.linspace(0, 2, 512))
z = (numpy.sin(X) + numpy.cos(Y)) ** 2 + 0.5
return z
i1 = generate_image()
i2 = generate_image()
display.image(i1, title='gradient')
# display.images([i2, i2, i2, i2], width=200, title='super fabio', labels=['a', 'b', 'c', 'd'])
data = []
for i in range(15):
data.append([i, random.random(), random.random() * 2])
win = display.plot(data, labels=[ 'position', 'a', 'b' ], title='progress')
for i in range(15, 25):
time.sleep(0.2)
data.append([i, random.random(), random.random() * 2])
display.plot(data, win=win)
import cv2
import display
display.set_port(9000)
def generate_image():
X, Y = np.meshgrid(np.linspace(0, np.pi, 512), np.linspace(0, 2, 512))
z = (np.sin(X) + np.cos(Y))**2 + 0.5
return z
i1 = generate_image()
i2 = generate_image()
display.image(i1, title='gradient')
# display.images([i2, i2, i2, i2], width=200, title='super fabio', labels=['a', 'b', 'c', 'd'])
data = []
for i in range(15):
data.append([i, random.random(), random.random() * 2])
win = display.plot(data, labels=['position', 'a', 'b'], title='progress')
for i in range(15, 25):
time.sleep(0.2)
data.append([i, random.random(), random.random() * 2])
display.plot(data, win=win)
im = cv2.imread('example.png')
+ inputs_names[0][:-4] \
+ '_3D' + time_info.strftime("_%H%M")
if not os.path.exists('./Trained/' + out_folder_name):
os.mkdir('./Trained/' + out_folder_name)
# load images
input_textures = [Image.open('./Textures/' + name) for name in inputs_names]
input_textures_torch = [
Variable(prep(img)).unsqueeze(0).cuda() for img in input_textures
]
# display images
if disp:
for i, img in enumerate(input_textures):
img_disp = numpy.asarray(img, dtype="int32")
display.image(img_disp,
win=['input' + str(i)],
title=['Input texture d' + str(i)])
#define layers, loss functions, weights and compute optimization target
loss_layers = ['r11', 'r21', 'r31', 'r41', 'r51']
loss_fns = [GramMSELoss()] * len(loss_layers)
loss_fns = [loss_fn.cuda() for loss_fn in loss_fns]
w = [1, 1, 1, 1, 1]
#compute optimization targets
targets = []
for img in input_textures_torch:
targets.append([GramMatrix()(f).detach() for f in vgg(img, loss_layers)])
# training parameters
slice_size = 128 # training slice resolution (best if same as examples)
print(f, " - ", len(eps), " len: ", len(eyes))
result.append(np.asarray(eps))
for ei in range(len(eps)):
x, y, w, h = eps[ei]
if w > 0 and h > 0:
print("eye @ ", eps[ei])
eye_img = image[y:y + h, x:x + w]
cv2.imwrite(f"{args.output}/eye_{ei}_{f}", eye_img)
# cv2.imwrite(args.output + "/final.png", image)
# print(type(image), image.shape)
if args.display:
destRGB = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
destRGB = cv2.flip(destRGB, 0)
display.image(destRGB, win='eye', title='eyes', width=800, height=450)
# cv2.imshow('Lets wear Glasses', image)
# cv2.waitKey()
# cv2.destroyAllWindows()
# net = BlurConv(3, 7, 1.5)
# print(net)
print(len(result))
result = np.asarray(result)
print(result.shape)
numlist = np.asarray(numlist)
