如何在Xcode中用Python读取.dcm文件？

Question

我正在尝试创建一个应用程序，用于查看和分析DICOM切片。我之前在MATLAB中做过这个应用，但MATLAB的工具不够强大，无法制作出真正好看的界面，而且3D效果也不好。因此，我花了很长时间尝试使用ITK和VTK在Xcode中构建一个应用，但一直没有成功。一天，我发现了一个叫xcodeproject PythonDicomDocument的项目——这个用Python写的项目可以读取和显示DICOM图像！我看过关于Python和Cocoa的教程，但仍然不太明白这个项目是怎么工作的——它有一个文件PythonDicomDocumentDocument.py:

from Foundation import *
from AppKit import *
from iiDicom import *

import objc


import dicom
import numpy
import Image


class PythonDicomDocumentDocument(NSDocument):
imageView = objc.IBOutlet('imageView')

def init(self):

    self = super(PythonDicomDocumentDocument, self).init()
    self.image = None
    return self

def windowNibName(self):

    return u"PythonDicomDocumentDocument"

def windowControllerDidLoadNib_(self, aController):
    super(PythonDicomDocumentDocument, self).windowControllerDidLoadNib_(aController)
    if self.image:
        self.imageView.setImageScaling_(NSScaleToFit)
        self.imageView.setImage_(self.image)


def dataOfType_error_(self, typeName, outError):
    return None

def readFromData_ofType_error_(self, data, typeName, outError):
    return NO

def readFromURL_ofType_error_(self, absoluteURL, typeName, outError):
    if absoluteURL.isFileURL():
        slice = iiDcmSlice.alloc().initWithDicomFileSlice_(absoluteURL.path())

        dicomImage = slice.sliceAsNSImage_context_(True, None)

        if dicomImage:
            self.image = dicomImage
                            #self.image = dicomImage

            return True, None

    return False, None

还有一个文件main.m:

**#import "<"Python/Python.h>**

**#import "<"Cocoa/Cocoa.h>**


int main(int argc, char *argv[])
{
    NSAutoreleasePool *pool = [[NSAutoreleasePool alloc] init];

    NSBundle *mainBundle = [NSBundle mainBundle];
    NSString *resourcePath = [mainBundle resourcePath];
    NSArray *pythonPathArray = [NSArray arrayWithObjects: resourcePath,               [resourcePath stringByAppendingPathComponent:@"PyObjC"],      @"/System/Library/Frameworks/Python.framework/Versions/Current/Extras/lib/python/", nil];

setenv("PYTHONPATH", [[pythonPathArray componentsJoinedByString:@":"] UTF8String], 1);

    NSArray *possibleMainExtensions = [NSArray arrayWithObjects: @"py", @"pyc",            @"pyo", nil];
    NSString *mainFilePath = nil;

for (NSString *possibleMainExtension in possibleMainExtensions) {
    mainFilePath = [mainBundle pathForResource: @"main" ofType: possibleMainExtension];
    if ( mainFilePath != nil ) break;
}

if ( !mainFilePath ) {
    [NSException raise: NSInternalInconsistencyException format: @"%s:%d main() Failed to find the Main.{py,pyc,pyo} file in the application wrapper's Resources directory.", __FILE__, __LINE__];
}

Py_SetProgramName("/usr/bin/python");
Py_Initialize();
PySys_SetArgv(argc, (char **)argv);

    const char *mainFilePathPtr = [mainFilePath UTF8String];

    FILE *mainFile = fopen(mainFilePathPtr, "r");



int result = PyRun_SimpleFile(mainFile, (char *)[[mainFilePath lastPathComponent] UTF8String]);



if ( result != 0 )
    [NSException raise: NSInternalInconsistencyException
                format: @"%s:%d main() PyRun_SimpleFile failed with file '%@'.  See console for errors.", __FILE__, __LINE__, mainFilePath];

[pool drain];

return result;

}

所以我想把MATLAB中读取.dcm文件的代码“翻译”过来：

directory = uigetdir; % after this command Finder window will appear and user will             choose a folder with .dcm files

fileFolder = directory; % the path to the folder is saved to a variable fileFolder
dirOutput = dir(fullfile(fileFolder,'*.dcm')); % choose files .dcm in specified folder     %and save their names 
fileNames = {dirOutput.name}';

Names = char(fileNames);
numFrames = numel(fileNames); % count the number of files in the folder


    for i = 1:numFrames
    Volume(:,:,i) = dicomread(fullfile(fileFolder,Names(i,:))); % create a 3D array of         %DICOM pixel data
    end;

有没有人能告诉我怎么在Xcode中用Python运行相同的代码来读取.dcm文件？？？

我听说Python和MATLAB很相似。

Answer 1

恭喜你选择Python来处理DICOM文件！根据我的经验，SciPy、numpy和matplotlib这几个库在处理大量数据时，比MATLAB（或者GNU Octave）要好得多。

下面是一些使用GDCM的Python接口、GDCM示例中的ConvertNumpy.py和matplotlib的加载和显示代码：

#!/usr/bin/env python

import gdcm
import ConvertNumpy
import numpy as np
import matplotlib.pyplot as plt

def loadDicomImage(filename):
    reader=gdcm.ImageReader()
    reader.SetFileName(filename)
    reader.Read()
    gdcmimage=reader.GetImage()
    return ConvertNumpy.gdcm_to_numpy(gdcmimage)

image=loadDicomImage('mydicomfile.dcm')

plt.gray()
plt.imshow(image)
plt.show()

需要注意的是，如果你的DICOM数据中有一些“填充”值远远超出了图像的空气和骨骼范围，可能会让imshow的自动缩放搞混；你可以使用vmax和vmin参数来指定你想要看到的范围，或者自己写一个简单的窗口级别调整代码（在numpy中这很简单）。