Abraham Ben

medpy with graphcut support

此时需要重新安装python。1. 从官网上下载tk和tcl下载tcl-tk2. 安装# install tclwget -c https://prdownloads.sourceforge.net/tcl/tcl8.6.9-src.tar.gztar -xvzf tcl8.6.9-src.tar.gzcd tcl8.6.9/unixmkdir buildcd build../configure --prefix=/home/xxx/local/tcl-tk/makemake ins

在程序的开头加上:import ctypeslibgcc_s = ctypes.CDLL('libgcc_s.so.1')即可解决

import sysorig_stdout = sys.stdoutf = open('out.txt', 'w')sys.stdout = ffor i in range(2): print('i = ', i)sys.stdout = orig_stdout # 这一句必须有，否则会报错f.close()

import cv2import matplotlib.pyplot as pltimg = cv2.imreadmulti("testsampleLV movie 2.tif", [], cv2.IMREAD_UNCHANGED)# img is a tuple, the first element is True, the second element is a list of array.plt.imshow(img[0][1])plt.show()

ii16 = np.iinfo(np.uint16)print(ii16.max, ii16.min)

使用joblib多线程1. 导包from joblib import Parallel, delayed2. 编写相关函数def func(arg1, arg2): res = do_something(arg1, arg2) return res3. 使用arg2 = ""# result is a listresults = Parallel(n_jobs=-1)( delayed(func)(arg1, arg2) for arg1 in tq

import numpy as npimport SimpleITK as sitkarr = np.array(...)img_itk = sitk.GetImageFromArray(arr)img_itk.SetSpacing(spacing)img_itk.SetOrigin(origin)img_itk.SetDirection(direction)sitk.WriteImage(img_itk, file_name.nii.gz)

python计算矩阵的connected components1. 使用scipy labelfrom scipy.ndimage import labelimport numpy as npfrom scipy.ndimage.morphology import generate_binary_structurearr = np.array([ [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], [0,

1. 新建一个vtkPolyDataimport vtkpoints = vtk.vtkPoints()points.InsertNextPoint(0, 0, 0)points.InsertNextPoint(1, 0, 0)points.InsertNextPoint(1, 1, 0)points.InsertNextPoint(0, 1, 0)points.InsertNextPoint(2, 0, 0)polygon = vtk.vtkPolygon() # 多边形polyg

1. 导包import numpy as npfrom os import pathfrom io import StringIOimport gzip2. 构建头部specific_header_keys = ("XDIM", "YDIM", "ZDIM", "VDIM", "TYPE", "PIXSIZE", "SCALE", "CPU", "VX", "VY", "VZ", "TX",

1. 导包from os import pathimport gzipfrom io import StringIOimport cv2import numpy as npimport matplotlib.pyplot as plt2. 读取文件filename = "skull_2.9.inr"if path.splitext(filename)[1] in (".gz", ".zip"): fzip = gzip.open(filename, "rb") f =

1. train_test_split进行一次性划分import numpy as npfrom sklearn.model_selection import train_test_splitX, y = np.arange(10).reshape((5, 2)), range(5)"""X: array([[0, 1], [2, 3], [4, 5], [6, 7], [8, 9]])lis

1. 加bound boxgray-scaleimg = cv2.rectangle(img, (point_1_x, point_1_y), (point_2_x, point_2_y), color=img.max(), thickness=1)plt.imshow(img, cmap="gray")如果想在gray-scale上加彩色bounding box:img = np.dstack([img_arr, img_arr, img_arr])*255 # img_arr为[0,

其实就是定义一个连通规则(structure)，用该连通区域在图像上stride, 用连通区域内的最小或者最大值代替原来的值。(边界上只覆盖部分值)1. 腐蚀(erosion)构造如下的连通区域:构造图像:img = np.array([ [125, 190, 11, 190], [141, 234, 21, 67], [165, 234, 31, 189], [112, 12, 41, 56] ], dtype='uint8')kernel = np

tqdm是python中打印进度条的一个简易工具包，可以方便查看循环的进度。具体见tqdm文档1. 搭配迭代器使用from tqdm import tqdmfor i in tqdm(range(10000)): passtqdm的构造函数为：tqdm(iterable=None, desc=None, total=None, leave=True, file=None, ncols=None, mininterval=0.1, maxinterval=10.0, miniters

1. setuptools的基本使用python的setuptools是用来将自己写的模块打包，然后生成.egg文件或者安装到site-packages目录下，然后就可以在文件中直接使用import来导入该包。使用方法如下:from setuptools import setup, find_packagessetup( name="pkg_name", version="1.0", packages=find_packages(), description="description",

安装某库conda install pkgname==version更新库conda update pkgname==version无法安装某包 则输入如下命令:anaconda serach -t conda pkgname得到结果为： 选择其中一个安装conda install -c https://conda.anaconda.org/AmirAlavi keras或者c

具体请见论文Greedy Function Approximation: A Gradient Boosting Machine1. GBDT的基本思想单模型情况下预测结果容易产生过拟合，例如普通决策树，要想达到比较好的预测效果，需要将树的深度调得比较深,叶节点的最大样本数目调得小一点等才能达到比较高的准确率。但是这样会带来严重的过拟合问题，针对这些问题，GBDT采用多颗决策树组合的方...

例如，使用清华大学镜像：pip install -i https://pypi.tuna.tsinghua.edu.cn/simple some-package如果要一次性一劳永逸解决, 修改 ~/.pip/pip.conf (没有就创建一个)， 修改 index-url至tuna，内容如下：[global]index-url = https://pypi.tuna.tsin...

1. opencv-python包opencv的像素值在[0,1][0,1]之间，存储的时候转换到[0,255][0,255]，show的时候转换到[0,255]import cv2img = cv2.imread("imgfile")cv2.imshow("img_win_name", img)cv2.waitKey(0) # 无限期等待输入cv2.imwrite("write_file_

1. 安装pycharm-debug.eggpycharm要实现远程调试，需要安装远程调试需要的安装包pycham-debug.egg(如果是python3,该文件为pycharm-debug-py3k.egg),通常位于pycharm安装目录下的”debug-eggs/”目录下。将该egg包拷贝到远程linux服务器上，使用easy_install来安装该包：easy_install pychar

import datetimed=datetime.date(2016,7,24) #新建一个日期，注意前面的0不能有print(d) => 20106-07-24d=datetime.date.today() #获取当前日期print(d)print(d.year, d.month, d.day()print(d.weekday) =>打印周几,此时monday是第0天, sunday 6