Tips for Python beginners to use Scikit-image examples for themselves 3 Write to a file

Tips for Python beginners to use the Scikit-image example for themselves

Tips for Python beginners to use the Scikit-image example for themselves 2 Process multiple files

Next, I will write the content for Python beginners to play with the example of Scikit-image with a slight modification.

Continue to Normalized Cut

Will be on the agenda.

In [Tips 2 for Python beginners to use the Scikit-image example for themselves, processing multiple files], I was able to process various images.

However, you may be dissatisfied because the result cannot be saved in the file as it is.

So I would like to give you some tips for saving to a file.

** OpenCV-Save image with Python **

cv2.imwrite("out.png ", img) OpenCV-Python Tutorials >> Getting Started with Images

** Save image as serial number file **

python

for  i in range(len(names)):
    img = cv2.imread(names[i])
    gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
    oname = "gray_%04d.png "  % i
    cv2.imwrite(oname, gray)

The function equivalent to the formatting of prinft () / sprinft () / fprintf () in C language is Python standard library / string format operation It is in.

** Operations related to directories and file names **

os.path.dirname() os.path.basename() os.path.splitext() os.path.join()

Python standard library / os.path — common path name operation

If the output directory odir does not exist, It is customary to create a directory as follows:

python

if not os.path.isdir(odir):
    os.makedirs(odir)

Python standard library / os — miscellaneous operating system interfaces

os.makedirs (path [, mode]) Recursive directory creation function. Similar to mkdir (), but creates all the intermediate directories needed to create the leaf directory.

oname = os.path.join(odir, "img_%05d.png " % i) If you include the output directory in the output file name like You can output to a directory other than the script execution directory.

** Operations related to string replacement **

The output file name can also be generated from the input file name by the replace function. The string has a replace () method, which returns the replaced string as a return value. replace(old, new[, count]) Python standard library / string methods

Using these functions, you can freely set the file name of the output destination.

Task:

"Hint: Let's process multiple files using for name in glob.glob (" * .jpg "):" Let's try to save the result to a file based on the example shown in. Condition: Create an output directory "out" and save the file there. Condition: For the input "image.jpg ", save it with the file name "out / image_ncut.png ".

You should be able to code with the hints given so far. Enjoy the OpenCV-Python, scikit-image, scikit-learn libraries even if you are new to Python.

** Note: **

-When writing directory delimiters in Python, you can use the "/" character like "abc / def.txt" even on Windows. -If you want to use the delimiter with "", it is convenient to add the symbol r before "" like name = r "abc \ def.txt". In that case, you don't have to write name = "abc \\ def.txt" (two "" characters are written).

In Python 3

https://docs.python.jp/3/library/pathlib.html Seems to be better to use.

python

# -*- coding: utf-8 -*-
# pylint: disable=C0103
import pathlib
from skimage import io, segmentation, color
from skimage.future import graph

def plotNcut(img):
    """
    img: color img

    """

    labels1 = segmentation.slic(img, compactness=30, n_segments=400)
    g = graph.rag_mean_color(img, labels1, mode='similarity')
    labels2 = graph.cut_normalized(labels1, g)
    return color.label2rgb(labels2, img, kind='avg'), labels2

if __name__ == "__main__":
    oDir = pathlib.Path("out")
    if not oDir.is_dir():
       oDir.mkdir() 
    for p in pathlib.Path("../images").glob("*.png "):
        print(p)
        
        oname = oDir.joinpath(p.with_suffix(".jpg ").name)
        img = io.imread(p)
        out, _ = plotNcut(img)
        io.imsave(str(oname), out)
        print(oname)

Hint 4 Use GUI