AttributeError: 'numpy.ndarray' object has no attribute 'getdraw'

Question

Okay so i'm trying to add an indian/nepali font for a gesture recognition system and i'm struggling a bit with the Pillow library. I followed the documentation for pillow and i'm getting an AttributeError.

My Code:

def put_splitted_text_in_blackboard(blackboard, splitted_text):
draw = ImageDraw.Draw(blackboard)
for text in splitted_text:
    fonts = ImageFont.truetype("preeti.TTF", 50)
    draw.text((10, 25), text, font=fonts)

It's giving me this error:

Traceback (most recent call last):
File "C:\Program Files\Python36\lib\site-packages\PIL\ImageDraw.py", line 
289, in Draw
return im.getdraw(mode)
AttributeError: 'numpy.ndarray' object has no attribute 'getdraw'

The blackboard has been defined here:

def recognize()
blackboard = np.zeros((480, 640, 3), dtype=np.uint8)
    splitted_text = split_sentence(text, 2)
    put_splitted_text_in_blackboard(blackboard, splitted_text)
    cv2.rectangle(img, (x, y), (x + w, y + h), (0, 255, 0), 2)
    res = np.hstack((img, blackboard))
    cv2.imshow("Recognizing gesture", res)
    cv2.imshow("thresh", thresh)
    if cv2.waitKey(1) == ord('q'):
        break

Anyone know where it went wrong?

The first arg to ImageDraw.Draw should be a PIL Image object, but you passed it a Numpy array. — PM 2Ring
– PM 2Ring, Commented Jun 17, 2018 at 10:45
okay yeah i got it, my bad but if i want to display characters on a blank screen, how do i pass a PIL Image to do so? — Shibesh Duwadi
– Shibesh Duwadi, Commented Jun 17, 2018 at 10:52
You can use PIL Image.new to create a blank image of whatever size you want, and a Image object has a .show method you can use to display it. There are fast methods to convert a PIL Image to a Numpy array, and vice versa, in case you need Numpy to do some processing that's not convenient in PIL. It's probably not a good idea to mix PIL & cv2 in the one program. It won't hurt anything, but it's a recipe for confusion. ;) For example, PIL uses the RGB convention, but cv2 uses BGR. — PM 2Ring
– PM 2Ring, Commented Jun 17, 2018 at 10:57
Like if i was just to use one of the pre-defined fonts in OpenCV i'd use: cv2.putText(blackboard, text, (4, 10), cv2.FONT_HERSHEY_SCRIPT_COMPLEX, 2, (255, 255, 255)) i'm still very much confused as to how PIL works. — Shibesh Duwadi
– Shibesh Duwadi, Commented Jun 17, 2018 at 11:00
In PIL you use ImageFont. It's probably a good idea at this stage to spend a bit of time browsing the PIL docs to get an idea of the tools that it offers. — PM 2Ring
– PM 2Ring, Commented Jun 17, 2018 at 11:02

Paula Thomas · Accepted Answer · 2018-06-17 11:24:07Z

7

What you need to do is change this line:

draw = ImageDraw.Draw(blackboard)

to

draw = ImageDraw.Draw(Image.fromarray(blackboard))

Which give ImageDraw an Image which it can understand not a numpy array which it can't.

edited Jun 17, 2018 at 11:24

answered Jun 17, 2018 at 11:12

Paula Thomas

1,1901 gold badge10 silver badges15 bronze badges

Sign up to request clarification or add additional context in comments.

4 Comments

Paula Thomas Over a year ago

Sorry typo: should be fromarray!!

Shibesh Duwadi Over a year ago

Thank you so much. That did the trick. Side question: any idea as to why openCV doesn't support international fonts? Kinda weird that it only supports one font family.

Paula Thomas Over a year ago

If you could accept my answer - I will edit to remove typo. Don't know OpenCV very well at all so no I doon't, maybe cotact development team via github?

Shibesh Duwadi Over a year ago

Done. Thanks for your help Paula.

Collectives™ on Stack Overflow

AttributeError: 'numpy.ndarray' object has no attribute 'getdraw'

1 Answer 1

4 Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

4 Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related