Skip to content
Load an image file into a numpy array with Exif orientation support. Prevents upside-down and sideways images!
Branch: master
Clone or download
Latest commit 3b353cf Oct 9, 2019
Type Name Latest commit message Commit time
Failed to load latest commit information.
image_to_numpy Initial commit Oct 9, 2019
tests Initial commit Oct 9, 2019
.gitignore Initial commit Oct 9, 2019
LICENSE Initial commit Oct 9, 2019 Update Oct 9, 2019 Initial commit Oct 9, 2019


Load an image file into a numpy array - while automatically rotating the image based on Exif orientation. Prevents upside-down and sideways images!

import image_to_numpy

img = image_to_numpy.load_image_file("my_file.jpg")

The image is automatically rotated into the correct orientation if the image contains Exif orientation metadata. Otherwise, it is loaded normally.

From there, you can pass the numpy array to any Python library that works with images in numpy array format, like face_recognition, Keras, etc.


You can install from PyPI:

pip install image_to_numpy

What is Exif Orientation data?

Most images captured by cell phones and consumer cameras aren't stored on disk in the same orientation they appear on screen. Exif Orientation data tells the program which way the image needs to be rotated to display correctly. Not handling Exif Orientation is a common source of bugs in Computer Vision and Machine Learning applications.

You can learn more about images and Exif Orientation data in my article here.


import image_to_numpy

img = image_to_numpy.load_image_file("my_file.jpg")

Your image is loaded - with the correct orientation!

By default, the image array is returned as a numpy array with 3-channels of 8-bit RGB data.

You can control the output format by passing in an optional mode parameter:

import image_to_numpy

img = image_to_numpy.load_image_file("my_file.jpg", mode="RGB")

# Supported modes:
#  1 (1-bit pixels, black and white, stored with one pixel per byte)
#  L (8-bit pixels, black and white)
#  RGB (3x8-bit pixels, true color)
#  RGBA (4x8-bit pixels, true color with transparency mask)
#  CMYK (4x8-bit pixels, color separation)
#  YCbCr (3x8-bit pixels, color video format)
#  I (32-bit signed integer pixels)
#  F (32-bit floating point pixels)

If you have matplotlib installed, here's a quick way to show your image on the screen:

import matplotlib.pyplot as plt
import image_to_numpy

img = image_to_numpy.load_image_file("my_file.jpg")



You can’t perform that action at this time.