Dlib: Support for reading images in correct orientation using EXIF data.

[RafayAK]

Current Behavior

I'm working on image processing with some images I collected myself. Dlib's dlib.load_rgb_image('image_path') method swaps the rows and columns on some images while OpenCV's cv2.imread('image_path') method does not.

Check out the results below

img = dlib.load_rgb_image("myimg.jpg")
print(img.shape)

--------------------
OUTPUT: (1944, 2592, 3)
(the resultant image is rotated 90 degrees clockwise)

while OpenCV's method returns the correct shape:

img = cv2.imread("myimg.jpg")
print(img.shape)

--------------------
OUTPUT: (2592, 1944, 3)

dlib.load_rgb_image() does not take into account the EXIF orientation metadata, so some images are read incorrectly.

I don't want to go in and rotate some of these offending images myself manually because I'm creating an app.

Is there a way in Dlib to read images using orientation information?

_Note: I asked this question of stackoverflow, one of the comments told me to create an issue here_

• Version: 19.17.0
• Where did you get dlib: pip
• Platform: Windows 10 - 64bit
• Compiler: python 3.6

[RafayAK]

Added platform info

[davisking]

Yeah, it doesn't do anything with EXIF data. It would be cool if the loader used it. Someone should submit a pull request that adds that feature :)

[RafayAK]

I'll see what I can do.

[nmaynes]

Would this require changes somewhere towards the top of image_loader.h?

[davisking]

That would be sensible.