def main(mode="random", iter_num=20, k=5):
"""average time to 1 pair = (2 / k^2) sec"""
if mode == "random":
from random import uniform
from time import time
img1 = cv2.imread("comp1.tif")
img1 = tools.resize(img1, k)
cont = []
for i in range(iter_num):
dx = dy = 0
dr = 0
shift = (uniform(-dx, dx), uniform(-dy, dy))
rot = (uniform(-dr, dr))
img2 = tools.get_img(img1, shift, rot)
solver = Solution(img1, img2, 100)
print("RIGHT", rot, shift)
start = time()
solver.estimate()
shift1 = [0, 0]
rot1, shift1[0], shift1[1] = solver.rotXY
if rot1 != False:
print("EST", rot1, shift1)
print("rotXY", solver.rotXY)
delta = [
abs(rot - rot1),
abs(shift[0] - shift1[0]),
abs(shift[1] - shift1[1])
]
print(f"FALLIBILITY = {delta}")
else:
print("NOT")
cont.append(time() - start)
print("___")
print(
f"MAX = {max(cont)}, MIN = {min(cont)}, AVERAGE = {sum(cont) / iter_num} SECONDS"
)
elif mode == "compare":
name1, name2 = f"/MESSIDOR/1pp.tif", f"/MESSIDOR/20051216_43913_0200_PP.tif"
img1 = (cv2.imread(name1))
img2 = (cv2.imread(name2))
#img2 = tools.rotAlignment(img2, 0)
#tools.show(img1)
#tools.show(img2)
img1 = segm.extract_bv(tools.resize(img1, k))
img2 = segm.extract_bv(tools.resize(img2, k))
solver = Solution(img1, img2)
print(solver.estimate())
return 0
def tell(path):
# print(model.to_json())
# import h5py as h5
# h = h5.File("./output/lenet.hdf5","r")
# print(h.keys())
# image = cv2.imread("./smiles/positives/positives7/24.jpg")
image = cv2.imread(path)
image = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
image = tools.resize(image, 28)
cv2.imwrite("./smiles/mine/mine_proc.jpg", image)
image = img_to_array(image)
data = []
data.append(image)
data = np.array(data, dtype="float") / 255.0
# print(np.info(data))
if not image.any():
print("No image input!")
else:
model = load_model('./output/lenet.hdf5')
predict = model.predict(data)
print(predict)
if predict.argmax(axis=1)[0]: # positive = 1， negative = 0
print("Positive!")
else:
print("Negative!")
def _generateCover(self, inFile, outFile, format, max_width, max_height):
from PIL import Image
try:
# Open the image
cover = Image.open(inFile)
width = cover.size[0]
height = cover.size[1]
newWidth, newHeight = tools.resize(width, height, max_width,
max_height)
cover = cover.resize((newWidth, newHeight), Image.ANTIALIAS)
cover.save(outFile, format)
except Exception:
# This message will probably be displayed for the thumbnail and the big cover.
logger.error(
'[%s] An error occurred while generating the cover for "%s"\n\n%s'
% (MOD_NAME, inFile, traceback.format_exc()))
# Remove corrupted file.
tools.remove(outFile)
# 命令行参数
ap = argparse.ArgumentParser()
ap.add_argument("-d", "--dataset", required=True)
ap.add_argument("-m", "--model", required=True)
args = vars(ap.parse_args())
# 数据和标签列表
data = []
labels = []
epoch = 10 # default: 15
# 遍历每一张样本图片
for imagePath in sorted(list(tools.list_images(args["dataset"]))):
image = cv2.imread(imagePath)
image = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
image = tools.resize(image, width=28)
image = img_to_array(image)
data.append(image)
# 样本图片路径smiles/positives/positives7/13229.jpg
label = imagePath.split(os.path.sep)[-3]
label = "smiling" if label == "positives" else "not_smiling"
labels.append(label)
# 缩放到[0,1]
# print(data[0][0])
data = np.array(data, dtype="float") / 255.0
# print(data[0][0])
# print(np.info(data))
labels = np.array(labels)
klick = pygame.mixer.Sound("klick.wav")
gong = pygame.mixer.Sound("gong.wav")
error = pygame.mixer.Sound("error.wav")
dinfo = pygame.display.Info()
video_flags = OPENGL | HWSURFACE | DOUBLEBUF | FULLSCREEN
#ein paar variablen zu den seitenverhaeltnissen:
SCREEN_RATIO = float(dinfo.current_w) / float(dinfo.current_h) #16:9
GITTER_RATIO = float(gitter.GITTER_W) / float(gitter.GITTER_H)
FIELD_W_RATIO = (GITTER_RATIO * float(dinfo.current_h)) / float(
dinfo.current_w)
pygame.display.set_mode((dinfo.current_w, dinfo.current_h), video_flags)
tools.resize((dinfo.current_w, dinfo.current_h))
tools.GlInit()
tools.clearScreen()
eManager = eventmanager.eventmanager()
mausimaus = maus.maus()
git = gitter.gitter((dinfo.current_w * FIELD_W_RATIO, dinfo.current_h),
eManager, mausimaus, (klick, gong, error))
sidebar = hud.hud(
(dinfo.current_w * FIELD_W_RATIO + 2, 2, dinfo.current_w -
dinfo.current_w * FIELD_W_RATIO - 4, dinfo.current_h - 4), mausimaus,
font)
import cv2
ap = argparse.ArgumentParser()
ap.add_argument("-c", "--cascade", required=True)
ap.add_argument("-m", "--model", required=True)
args = vars(ap.parse_args())
detector = cv2.CascadeClassifier(args["cascade"])
model = load_model(args["model"])
camera = cv2.VideoCapture(0)
while True:
(grabbed, frame) = camera.read()
frame = tools.resize(frame, width=900)
gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
frameNew = frame.copy()
# 在灰度图上检测人脸，得到所有矩形。
rects = detector.detectMultiScale(gray,
scaleFactor=1.1,
minNeighbors=5,
minSize=(30, 30),
flags=cv2.CASCADE_SCALE_IMAGE)
# 循环每一个检测出人脸的矩形区域
for (fX, fY, fW, fH) in rects:
# 抽取ROI区域，缩放为28*28
roi = gray[fY:fY + fH, fX:fX + fW]
roi = cv2.resize(roi, (28, 28))
multinom_bkg = list(np.random.multinomial(1, prob_bkg))
draw_bkg = multinom_bkg.index(1)
bkg_name = background_list[draw_bkg]
multinom_card = list((np.random.multinomial(1, prob_card)))
draw_card = multinom_card.index(1)
card_name = card_list[draw_card]
bkg = PIL.Image.open(path + "/Backgrounds/" + bkg_name)
card = Image.open(path + "/Cards/" + card_name)
card = card.convert('RGBA')
# Give random size on card image
ratio = np.random.normal(2.2, 0.1)
print("Card ratio: " + str(round(ratio, 2)))
card = tools.resize(card, ratio)
# Set a paste position of card around the middle of the background image.
# Top left corner position is starting coords (0,0), not middle.
x_val = bkg.size[0] / 2 - card.size[0] / 2
y_val = bkg.size[1] / 2 - card.size[1] / 2
x_offset = int(np.random.normal(x_val, x_val / 5.5))
y_offset = int(np.random.normal(y_val, y_val / 5.5))
# Copy images
back_im = bkg.copy()
card_im = card.copy()
# Set variables to be used in mutation
w, h = card_im.size
checkpoint = torch.load('../fast-depth/mobilenet-nnconv5dw-skipadd.pth.tar',
map_location=torch.device('cpu'))
torch_model = checkpoint['model']
import cv2
import tools
cap = cv2.VideoCapture('/home/sebastian/Desktop/SLAM-CPP/home.MOV')
ret, frame = cap.read()
while False:
#frame = cv2.resize(frame, (224, 224))
frame = tools.resize(frame)
cv2.imshow('frame', frame)
model_in = np.expand_dims(np.array(frame, dtype=np.float32)/255., 0)
tf_prediction = model.predict(model_in)
tf_prediction = np.squeeze(tf_prediction)
tf_prediction = tools.toShow(tf_prediction)
model_in = np.moveaxis(model_in, -1, 1)
torch_model_in = torch.tensor(model_in).float()
torch_prediction = torch_model(torch_model_in)
torch_prediction = torch_prediction.detach().numpy().squeeze()
torch_prediction = tools.toShow(torch_prediction)
cv2.imshow('tf_prediction', tf_prediction)
cv2.imshow('torch_prediction', torch_prediction)