Importing PNG files into Numpy? - python

I have about 200 grayscale PNG images stored within a directory like this.
1.png
2.png
3.png
...
...
200.png
I want to import all the PNG images as NumPy arrays.
How can I do this?

According to the doc, scipy.misc.imread is deprecated starting SciPy 1.0.0, and will be removed in 1.2.0. Consider using imageio.imread instead.
Example:
import imageio
im = imageio.imread('my_image.png')
print(im.shape)
You can also use imageio to load from fancy sources:
im = imageio.imread('http://upload.wikimedia.org/wikipedia/commons/d/de/Wikipedia_Logo_1.0.png')
Edit:
To load all of the *.png files in a specific folder, you could use the glob package:
import imageio
import glob
for im_path in glob.glob("path/to/folder/*.png"):
im = imageio.imread(im_path)
print(im.shape)
# do whatever with the image here

Using just scipy, glob and having PIL installed (pip install pillow) you can use scipy's imread method:
from scipy import misc
import glob
for image_path in glob.glob("/home/adam/*.png"):
image = misc.imread(image_path)
print image.shape
print image.dtype
UPDATE
According to the doc, scipy.misc.imread is deprecated starting SciPy 1.0.0, and will be removed in 1.2.0. Consider using imageio.imread instead. See the answer by Charles.

This can also be done with the Image class of the PIL library:
from PIL import Image
import numpy as np
im_frame = Image.open(path_to_file + 'file.png')
np_frame = np.array(im_frame.getdata())
Note: The .getdata() might not be needed - np.array(im_frame) should also work

Using a (very) commonly used package is prefered:
import matplotlib.pyplot as plt
im = plt.imread('image.png')

If you are loading images, you are likely going to be working with one or both of matplotlib and opencv to manipulate and view the images.
For this reason, I tend to use their image readers and append those to lists, from which I make a NumPy array.
import os
import matplotlib.pyplot as plt
import cv2
import numpy as np
# Get the file paths
im_files = os.listdir('path/to/files/')
# imagine we only want to load PNG files (or JPEG or whatever...)
EXTENSION = '.png'
# Load using matplotlib
images_plt = [plt.imread(f) for f in im_files if f.endswith(EXTENSION)]
# convert your lists into a numpy array of size (N, H, W, C)
images = np.array(images_plt)
# Load using opencv
images_cv = [cv2.imread(f) for f in im_files if f.endswith(EXTENSION)]
# convert your lists into a numpy array of size (N, C, H, W)
images = np.array(images_cv)
The only difference to be aware of is the following:
opencv loads channels first
matplotlib loads channels last.
So a single image that is 256*256 in size would produce matrices of size (3, 256, 256) with opencv and (256, 256, 3) using matplotlib.

To read in one image:
import PIL.Image
im = PIL.Image.open('path/to/your/image')
im = np.array(im)
Iterate to read in multiple images.
This answer is similar to this but simpler (no need for .getdata()).

I changed a bit and it worked like this, dumped into one single array, provided all the images are of same dimensions.
png = []
for image_path in glob.glob("./train/*.png"):
png.append(misc.imread(image_path))
im = np.asarray(png)
print 'Importing done...', im.shape

I like the build-in pathlib libary because of quick options like directory= Path.cwd()
Together with opencv it's quite easy to read pngs to numpy arrays.
In this example you can even check the prefix of the image.
from pathlib import Path
import cv2
prefix = "p00"
suffix = ".png"
directory= Path.cwd()
file_names= [subp.name for subp in directory.rglob('*') if (prefix in subp.name) & (suffix == subp.suffix)]
file_names.sort()
print(file_names)
all_frames= []
for file_name in file_names:
file_path = str(directory / file_name)
frame=cv2.imread(file_path)
all_frames.append(frame)
print(type(all_frames[0]))
print(all_frames[0] [1][1])
Output:
['p000.png', 'p001.png', 'p002.png', 'p003.png', 'p004.png', 'p005.png', 'p006.png', 'p007.png', 'p008.png', 'p009.png']
<class 'numpy.ndarray'>
[255 255 255]

If you prefer the standard library:
#IMPORTANT: This Code only works with Python>=3.6
Directory="."#Your directory
import os
import tkinter
import numpy
tk=tkinter.Tk()
tk.overrideredirect(1)
tk.geometry("0x0")
Items=[]
for i in os.listdir(Directory):
fn=Directory+os.sep+i
imgArray=[]
image=tkinter.PhotoImage(file=fn)
for w in range(image.width()):
entry=[]
for h in range(image.height()):
entry.append(image.get(w,h))
imgArray.append(entry)
imgArray=numpy.array(imgArray)
Items.append(imgArray)
tk.destroy()

Related

How to load images from memory to numpy using file system

I want to store my image directory in memory, then load the images into a numpy array.
The normative way to load images that are not in memory is as follows:
import PIL.Image
import numpy as np
image = PIL.Image.open("./image_dir/my_image_1.jpg")
image = np.array(image)
However, I am not sure how to do this when the images are in memory. So far, I have been able to setup the following starter code:
import fs
import fs.memoryfs
import fs.osfs
image_dir = "./image_dir"
mem_fs = fs.memoryfs.MemoryFS()
drv_fs = fs.osfs.OSFS(image_path)
fs.copy.copy_fs(drv_fs, mem_fs)
print(mem_fs.listdir('.'))
Returns:
['my_image_1.jpg', 'my_image_2.jpg']
How do I load images that are in memory into numpy?
I am also open to alternatives to the fs package.
As per the documentation, Pillow's Image.open accepts a file object instead of a file name, so as long as your in-memory file package provides Python file objects (which it most likely does), you can just use them. If it doesn't, you could even just wrap them in a class that provides the required methods. Assuming you are using PyFilesystem, according to its documentation you should be fine.
So, you want something like:
import numpy as np
import PIL.Image
import fs.memoryfs
import fs.osfs
import fs.copy
mem_fs = fs.memoryfs.MemoryFS()
drv_fs = fs.osfs.OSFS("./image_dir")
fs.copy.copy_file(drv_fs, './my_image_1.jpg', mem_fs, 'test.jpg')
with mem_fs.openbin('test.jpg') as f:
image = PIL.Image.open(f)
image = np.array(image)
(note I just used copy_file because I tested with a single file, you can use copy_fs if you need to copy the entire tree - it's the same principle)

Working with TIFF in Python using EasyOCR

I use a Python Module EasyOCR for extracting text from image. This Method works for PNG Format but in TIFF Situation give me a error
Code look like this:
import easyocr
import cv2
from matplotlib import pyplot as plt
import numpy as np
IMAGE_PATH = 'IMG_4022.tif'
reader = easyocr.Reader(['en'], gpu=False)
result = reader.readtext(IMAGE_PATH)
result
I work with Juypter Notebook
You are not reading the image. Please use opencv to read the image. Ensure that the image is in the current directory or provide the absolute path of the image.
from easyocr import Reader
import cv2
from matplotlib import pyplot as plt
import numpy as np
IMAGE_PATH = "IMG_4022.tif"
image = cv2.imread(IMAGE_PATH)
languages = ['en']
reader = Reader(languages, gpu = False)
results = reader.readtext(image)

Loading non RGBA 4-tile multi page tiff images with skimage into dask array

I need to read a folder full of multi-page tiffs generated by the Suite2p neurobiology package.
From the Suite2p source code the multi-tiffs are created as follows:
import numpy as np
from tifffile import TiffWriter
# fake example images
img_5_tiles = np.random.randint(0,65535,(5,10,20), dtype='uint16')
img_4_tiles = img_5_tiles[1:,...]
# save fake images
with TiffWriter(r'D:\5tiles.tiff') as tif:
for frame in img_5_tiles:
tif.save(frame)
with TiffWriter(r'D:\4tiles.tiff') as tif:
for frame in img_4_tiles:
tif.save(frame)
When I try to read them into dask, skimage.io default tifffile plugin fails to get the correct shape:
from dask.array.image import imread
from skimage import io
def read_with_tifffile(path):
return io.imread(path, plugin='tifffile')
# should produce (5,10,20) instead of (1,10,20)
imread(r'D:\5tiles.tiff', imread=read_with_tifffile).shape
I can overcome this by using the non-default skimage.io.imread plugin 'pil'
def read_with_pil(path):
return io.imread(path, plugin='pil', mode='L')
# gives (1,5,10,20) which is acceptable
imread(r'D:\5tiles.tiff', imread=read_with_pil).shape
Unfortunately, if the number of tiles equals to 4, skimage starts to handle the shape differently:
# gives (1,10,20,4) instead of (1,4,10,20)
imread(r'D:\4tiles.tiff', imread=read_with_pil).shape
From reading skimage docs, it's probably trying to interpret my image as RGBA and then fails.
Is there a solution to force 'uint16' multipage read for all image shapes?
Any advice is appreciated!
Based on cgohlke's comment:
from dask.array.image import imread
import numpy as np
from skimage import io
import tifffile
# fake data
img_with_4_tiles = np.random.randint(0,65535,(4,10,20), dtype='uint16')
with tifffile.TiffWriter(r'D:\4tiles.tiff') as tif:
for frame in img_with_4_tiles:
tif.save(frame)
# my original bad solution
def read_with_pil(path):
return io.imread(path, plugin='pil', mode='L')
print(imread(r'D:\4tiles.tiff', imread=read_with_pil).shape)
# a good solution from the horse's mouth (https://pypi.org/user/cgohlke/)
def read_with_tifffile(path):
return tifffile.imread(path, is_shaped=False)
print(imread(r'D:\4tiles.tiff', imread=read_with_tifffile).shape)

Importing PNG for deep-learning Python

I am a novice coder, having been self-taught through codeacademy. I am wondering what is the easiest way to import png files to python (2.7.14) with the goal of using these files in a deep-learning program.
So far I have tried these two codes:
import scipy
from scipy import misc
import glob
for image_path in glob.glob("/E:/_SAMM_Projects/gemini_hikai_DM_hack_complete/export/contact_frames/boat/*.png"):
image = misc.imread(image_path)
print image.shape
print image.dtype
import scipy
from scipy import misc
import glob
import numpy
png = []
for image_path in glob.glob("/E:/_SAMM_Projects/gemini_hikai_DM_hack_complete/export/contact_frames/boat/*.png"):
png.append(misc.imread(image_path))
im = np.asarray(png)
print "importing done...", im.shape
based off templates I have found online, both do not seem to work
In the context of deep learning, I understand that you would like to read an image into a numpy array so you can use deep learning models (such as ConvNet) on it.
I suggest using OpenCV for your purpose.
import cv2
image = cv2.imread("yourimg.png")

Viewing .npy images

How can I view images stored with a .npy extension and save my own files in that format?
.npy is the file extension for numpy arrays - you can read them using numpy.load:
import numpy as np
img_array = np.load('filename.npy')
One of the easiest ways to view them is using matplotlib's imshow function:
from matplotlib import pyplot as plt
plt.imshow(img_array, cmap='gray')
plt.show()
You could also use PIL or pillow:
from PIL import Image
im = Image.fromarray(img_array)
# this might fail if `img_array` contains a data type that is not supported by PIL,
# in which case you could try casting it to a different dtype e.g.:
# im = Image.fromarray(img_array.astype(np.uint8))
im.show()
These functions aren't part of the Python standard library, so you may need to install matplotlib and/or PIL/pillow if you haven't already. I'm also assuming that the files are either 2D [rows, cols] (black and white) or 3D [rows, cols, rgb(a)] (color) arrays of pixel values.
Thanks Ali_m. In my case I inspect the npy file to check how many images was in the file with:
from PIL import Image
import numpy as np
data = np.load('imgs.npy')
data.shape
then I plotted the images in a loop:
from matplotlib import pyplot as plt
for i in range(len(data)):
plt.imshow(data[i], cmap='gray')
plt.show()

Categories

Resources