def visualize_ori(cid, model_path):
seg = load_segmentation(cid)
visualize(cid,
seg,
os.path.join(model_path, 'pics',
str(cid).zfill(5) + '_gt'),
ori=True)
def visualize_case(case_result, model_path, case_id):
visualize(case_id,
case_result,
os.path.join(model_path, 'pics',
str(case_id).zfill(5) + '_pred'),
ori=False,
k_color=MAGENTA_RGB,
t_color=CYAN_RGB)
from pathlib import Path
import os.path
import sys
sys.path.append(os.path.join(os.path.dirname(__file__), '..'))
from configs import config
from tqdm import tqdm
from starter_code import utils
from starter_code import visualize
DATAPATH = config.ORIGINAL_DATA
if __name__ == '__main__':
for idx in tqdm(range(210)):
visualize.visualize(
idx,
DATAPATH + 'visualization-perspectives/case_{:05d}'.format(idx),
data_path=Path(DATAPATH),
separate=True,
plane='all')
# Código para el proyecto de imágenes
# Este starter code lo proporciona el mismo repositorio del challenge
import nibabel as nib
from starter_code.utils import load_case
from starter_code.visualize import visualize
volume, segmentation = load_case(0)
visualize("case_00003", im1)
from win32com.client import Dispatch
o = Dispatch("ThunderAgent.Agent64.1")
for i in range(210):
url1 = 'https://media.githubusercontent.com/media/neheller/kits19/master/data/case_%05d/imaging.nii.gz' % i
url2 = 'https://media.githubusercontent.com/media/neheller/kits19/master/data/case_%05d/segmentation.nii.gz' % i
filename1 = 'imaging_%03d.nii.gz' % i
filename2 = 'segmentation_%03d.nii.gz' % i
o.AddTask(url1, filename1)
o.AddTask(url2, filename2)
o.CommitTasks()
# %%
def visualize_from_pred(cid, model_path, path='./predictions'):
file_name = 'prediction_00' + str(cid) + '.nii.gz'
pred = nib.load(os.path.join(path, file_name))
pred = pred.get_fdata().copy()
visualize(cid, pred, os.path.join(model_path, 'pics'), ori=False)
def GenerateEvaluationpng(input_path, output_path):
for i in range(50):
name = "case_{:05d}".format(i + 160)
visualize(name, input_path, output_path)
def GenerateTrainpng(input_path, output_path):
for i in range(160):
name = "case_{:05d}".format(i)
visualize(name, input_path, output_path)
seg_data[seg_data == 2] = 1
a = int(math.ceil(len(pred_seg) / 48))
pred_seg = np.reshape(pred_seg, (a, 48, 128, 128, 1))
pred_seg = concat_3Dmatrices(pred_seg,
image_size=vol_data.shape,
window=(48, 128, 128),
overlap=(12, 32, 32))
if not os.path.exists("predictions"):
os.mkdir("predictions")
nifti = nib.Nifti1Image(pred_seg, None)
nib.save(
nifti,
os.path.join("predictions",
"prediction_" + str(cases[0]).zfill(5) + ".nii.gz"))
pred_seg = pred_seg[:, :, :, 0]
pred_seg[pred_seg <= 0.5] = 0
pred_seg[pred_seg > 0.5] = 1
visualize(vol_data, pred_seg, "evaluation")
visualize(vol_data, seg_data, "evaluation")
from preprocessing import preprocessing
from find_and_save_kidneys import find_and_save_kidneys
from save_to_png import save_to_png
import re
from skimage.transform import resize
from get_new_seeds import get_new_seeds
case_num = input("Please enter a case number:\n")
data_path = "case_" + case_num.zfill(5)
image_path = "case" + case_num
out_path = Path(image_path)
if not out_path.exists():
visualize(data_path, image_path,
t_color=[0, 255,
0]) #konwersja danych z repozytorium do obrazów .png
image, name, colour = find_image_with_kidneys(image_path + '/*[0-9].png')
image_rgb = io.imread(name)
image = (color.rgb2gray(image_rgb))
p = '[\d]+[\d]+[\d]+[\d]+[\d]'
if re.search(p, name) is not None:
for catch in re.finditer(p, name):
number = int(catch[0])
image_clear, crop = preprocessing(image)
image_rectangle = np.zeros((512, 512))
image_rectangle[200:450, 80:231] = 1
#zmiana working directory na folder zawierajacy zdjecie
import os
os.chdir('/Users/komala/Desktop/TOM_projekt/kits19')
from starter_code.utils import load_case
from starter_code.visualize import visualize
volume, segmentation = load_case("case_00000")
visualize("case_00000", "pictures")