When we are building web services using Python, we often send or receive images in base64 encoded format. However, when we are doing image processing tasks, we need to use PIL or OpenCV. In this post, I will share how to convert between OpenCV or PIL image and base64 encoded image.

base64 to PIL Image#

import base64
from io import BytesIO
from PIL import Image

with open("test.jpg", "rb") as f:
    im_b64 = base64.b64encode(f.read())

im_bytes = base64.b64decode(im_b64)   # im_bytes is a binary image
im_file = BytesIO(im_bytes)  # convert image to file-like object
img = Image.open(im_file)   # img is now PIL Image object

In the above code, since Image.open() only accepts image path or file-like object, we first convert the base64 encoded image to BytesIO object and then read the image using PIL.

base64 to OpenCV Image#

import base64
import numpy as np
import cv2

with open("test.jpg", "rb") as f:
    im_b64 = base64.b64encode(f.read())

im_bytes = base64.b64decode(im_b64)
im_arr = np.frombuffer(im_bytes, dtype=np.uint8)  # im_arr is one-dim Numpy array
img = cv2.imdecode(im_arr, flags=cv2.IMREAD_COLOR)

In the above code, we first convert binary image to Numpy array, then decode the array with cv2.imdecode(). The final img is an OpenCV image in Numpy ndarray format.

PIL or OpenCV image to base64#

PIL Image to base64#

import base64
from io import BytesIO
from PIL import Image

img = Image.open('test.jpg')
im_file = BytesIO()
img.save(im_file, format="JPEG")
im_bytes = im_file.getvalue()  # im_bytes: image in binary format.
im_b64 = base64.b64encode(im_bytes)

In the above code, instead of saving the PIL Image object img to the disk, we save it to im_file which is a file-like object. Note that in this case, we need to specify the image format in img.save().

OpenCV to base64 image#

import base64
import numpy as np
import cv2

img = cv2.imread('test.jpg')
_, im_arr = cv2.imencode('.jpg', img)  # im_arr: image in Numpy one-dim array format.
im_bytes = im_arr.tobytes()
im_b64 = base64.b64encode(im_bytes)

In the above code, we first save the image in Numpy ndarray format to im_arr which is a one-dim Numpy array. We then get the image in binary format by using the tobytes() method of this array.

References#

• Convert OpenCV or PIL image to bytes.
• base64 image to PIL Image.
• Byte array to OpenCV image.
• OpenCV image to base64.