Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

ImageDraw.polygon produces the wrong image #8616

Open
SakuMyl opened this issue Dec 22, 2024 · 5 comments · Fixed by #8620 · May be fixed by #8659
Open

ImageDraw.polygon produces the wrong image #8616

SakuMyl opened this issue Dec 22, 2024 · 5 comments · Fixed by #8620 · May be fixed by #8659

Comments

@SakuMyl
Copy link

SakuMyl commented Dec 22, 2024

What did you do?

Drew a polygon with PIL's Image and ImageDraw modules

What did you expect to happen?

The Polygon should look consistent without holes in the middle

What actually happened?

There are holes and gaps.

What are your OS, Python and Pillow versions?

  • OS: Ubuntu 22.04.5 LTS
  • Python: Python 3.10.12
  • Pillow: 11.0.0
Pillow 11.0.0
Python 3.10.12 (main, Nov  6 2024, 20:22:13) [GCC 11.4.0]
--------------------------------------------------------------------
Python executable is /home/ghrunner/venv/bin/python3
Environment Python files loaded from /home/ghrunner/venv
System Python files loaded from /usr
--------------------------------------------------------------------
Python Pillow modules loaded from /home/ghrunner/venv/lib/python3.10/site-packages/PIL
Binary Pillow modules loaded from /home/ghrunner/venv/lib/python3.10/site-packages/PIL
--------------------------------------------------------------------
--- PIL CORE support ok, compiled for 11.0.0
*** TKINTER support not installed
--- FREETYPE2 support ok, loaded 2.13.2
--- LITTLECMS2 support ok, loaded 2.16
--- WEBP support ok, loaded 1.4.0
--- JPEG support ok, compiled for libjpeg-turbo 3.0.4
--- OPENJPEG (JPEG2000) support ok, loaded 2.5.2
--- ZLIB (PNG/ZIP) support ok, loaded 1.2.11
--- LIBTIFF support ok, loaded 4.6.0
--- RAQM (Bidirectional Text) support ok, loaded 0.10.1, fribidi 1.0.8, harfbuzz 10.0.1
*** LIBIMAGEQUANT (Quantization method) support not installed
--- XCB (X protocol) support ok

Code to reproduce the issue along with the images:

from PIL import Image, ImageDraw
import matplotlib.pyplot as plt

shape = (35, 15)
pts = [(31, 12), (30, 12), (28, 11), (28, 10), (26, 7), (23, 7), (22, 8), (22, 13), (22, 13), (21, 13), (20, 14), (18, 14), (17, 13), (15, 13), (13, 11), (10, 11), (8, 13), (5, 13), (4, 12), (3, 12), (1, 10), (1, 8), (3, 5), (4, 6), (5, 6), (6, 7), (6, 8), (8, 10), (9, 10), (10, 10), (10, 4), (13, 0), (19, 0), (21, 2), (21, 3), (22, 4), (23, 3), (24, 3), (26, 6), (26, 7), (31, 12)]
img = Image.new("L", shape, 0)
draw = ImageDraw.Draw(img)
draw.polygon(pts, fill=255)
img.save('polygon.png')

x, y = zip(*pts)
plt.figure(figsize=shape)
plt.plot(x, y, marker='o', linestyle='-', color='blue', markersize=5, label='Points and Lines')

plt.scatter(x, y, color='red', zorder=5, label='Points')

plt.title("Visualization of Points with Connections")
plt.xlabel("X-axis")
plt.ylabel("Y-axis")
plt.grid(True)
plt.legend()
plt.savefig('points.png')

image
image
@radarhere radarhere mentioned this issue Dec 23, 2024
Merged
@radarhere
Copy link
Member

I've created #8620 to resolve this. With it, Pillow creates this image.

polygon

@SakuMyl
Copy link
Author

SakuMyl commented Jan 2, 2025 •
edited by radarhere
Loading

I think the bug fix might have created new issues. Take a look at this example:

from PIL import Image, ImageDraw
import matplotlib.pyplot as plt

pts = [(183.5, 158.0), (184.0, 157.5), (185.0, 157.5), (186.0, 156.5), (189.0, 156.5), (190.00000390625002, 155.5), (192.0, 155.5), (192.99999609374998, 154.5), (195.0, 154.5), (196.0, 153.5), (196.99999609374998, 153.5), (198.00000390625002, 152.5), (199.0, 152.5), (200.0, 151.5), (200.99999609374998, 151.5), (202.00000390625002, 150.5), (203.0, 150.5), (204.99999609374998, 148.5), (206.00000390625002, 148.5), (207.5, 147.0), (207.5, 146.0), (208.49999609374998, 145.0), (208.49999609374998, 144.0), (209.49999609374998, 143.0), (209.49999609374998, 142.0), (210.50000390625002, 141.0), (210.50000390625002, 140.0), (211.5, 139.0), (211.5, 138.0), (211.0, 137.5), (207.0, 141.5), (206.00000390625002, 141.5), (205.50000390625002, 142.0), (205.50000390625002, 143.0), (204.49999609374998, 144.0), (204.49999609374998, 145.0), (202.00000390625002, 147.5), (201.49999609374998, 147.0), (201.49999609374998, 146.0), (202.50000390625002, 145.0), (202.00000390625002, 144.5), (200.99999609374998, 144.5), (200.49999609374998, 144.0), (200.49999609374998, 143.0), (200.0, 142.5), (199.0, 142.5), (198.50000390625002, 142.0), (198.50000390625002, 141.0), (202.50000390625002, 137.0), (202.50000390625002, 136.0), (203.5, 135.0), (203.5, 133.0), (204.99999609374998, 131.5), (206.00000390625002, 131.5), (207.0, 130.5), (208.0, 130.5), (208.99999609374998, 129.5), (211.0, 129.5), (212.0, 128.5), (218.00000390625002, 128.5), (218.50000390625002, 129.0), (218.50000390625002, 130.0), (213.50000390625002, 135.0), (214.00000390625002, 135.5), (215.0, 134.5), (216.0, 134.5), (216.99999609374998, 133.5), (218.00000390625002, 133.5), (219.0, 132.5), (220.0, 132.5), (220.99999609374998, 131.5), (222.00000390625002, 131.5), (223.000001953125, 130.5), (223.999998046875, 130.5), (226.00000390625002, 128.5), (227.000001953125, 128.5), (227.999998046875, 127.5), (232.99999609374998, 127.5), (234.00000390625002, 126.5), (235.000001953125, 126.5), (237.500005859375, 124.0), (235.999998046875, 122.5), (235.000001953125, 122.5), (234.500001953125, 123.0), (234.500001953125, 124.0), (232.99999609374998, 125.5), (230.00000390625002, 125.5), (229.500005859375, 125.0), (232.499998046875, 122.0), (231.999998046875, 121.5), (230.00000390625002, 121.5), (228.99999609374998, 122.5), (224.99999609374998, 122.5), (223.999998046875, 123.5), (220.99999609374998, 123.5), (220.0, 124.5), (216.99999609374998, 124.5), (216.0, 125.5), (212.99999609374998, 125.5), (212.0, 126.5), (208.99999609374998, 126.5), (208.0, 127.5), (206.00000390625002, 127.5), (204.99999609374998, 128.5), (203.0, 128.5), (202.00000390625002, 129.5), (200.99999609374998, 129.5), (190.00000390625002, 140.5), (189.0, 140.5), (187.0, 142.5), (186.0, 142.5), (185.0, 143.5), (184.0, 143.5), (182.0, 145.5), (181.0, 145.5), (178.0, 148.5), (177.0, 148.5), (172.5, 153.0), (173.0, 153.5), (174.0, 152.5), (175.0, 152.5), (176.0, 151.5), (177.0, 151.5), (177.4999921875, 152.0), (177.4999921875, 154.0), (179.0, 155.5), (181.0, 155.5), (182.5, 154.0), (182.5, 151.0), (181.5000078125, 150.0), (181.5000078125, 149.0), (182.0, 148.5), (183.0, 149.5), (184.0, 149.5), (186.0, 151.5), (188.0, 151.5), (188.5, 151.0), (188.5, 150.0), (187.5, 149.0), (187.5, 147.0), (185.4999921875, 145.0), (186.0, 144.5), (188.0, 144.5), (188.5, 145.0), (188.5, 146.0), (189.0, 146.5), (192.0, 146.5), (194.00000390625002, 144.5), (195.0, 144.5), (197.50000390625002, 147.0), (194.00000390625002, 150.5), (192.99999609374998, 150.5), (191.5, 152.0), (191.5, 153.0), (191.0, 153.5), (189.0, 153.5), (188.0, 154.5), (187.0, 154.5), (185.0, 156.5), (181.0, 156.5), (180.0, 157.5), (179.0, 157.5), (178.5, 158.0), (179.0, 158.5), (183.0, 158.5)]
pts = [(round(x), round(y)) for x, y in pts]
shape = (max([x for x, y in pts]) + 1, max([y for x, y in pts]) + 1)
img = Image.new("L", shape, 0)
draw = ImageDraw.Draw(img)
draw.polygon(pts, fill=255)
img.save('polygon.png')

x, y = zip(*pts)
plt.figure(figsize=shape)
plt.plot(x, y, marker='o', linestyle='-', color='blue', markersize=5, label='Points and Lines')

plt.scatter(x, y, color='red', zorder=5, label='Points')

plt.title("Visualization of Points with Connections")
plt.xlabel("X-axis")
plt.ylabel("Y-axis")
plt.grid(True)
plt.legend()
plt.savefig('points.png')`

points
polygon

There is a strip of pixels added to the mask which should not be there

@radarhere radarhere reopened this Jan 2, 2025
@SakuMyl
Copy link
Author

SakuMyl commented Jan 2, 2025 •
edited by radarhere
Loading

Yet another example for debugging:

from PIL import Image, ImageDraw
import matplotlib.pyplot as plt

pts =  [(154.5, 189.0), (155.0, 188.5), (157.0, 188.5), (158.0, 187.5), (159.0, 187.5), (159.5, 187.0), (159.5, 186.0), (160.5, 185.0), (160.5, 182.0), (161.0, 181.5), (169.0, 181.5), (170.0, 182.5), (172.0, 180.5), (184.0, 180.5), (185.0, 179.5), (187.0, 179.5), (188.0, 178.5), (192.0, 178.5), (192.99999609374998, 177.5), (203.0, 177.5), (204.0, 176.5), (212.0, 176.5), (212.99999609374998, 177.5), (214.00000390625002, 177.5), (218.50000390625002, 182.0), (218.50000390625002, 183.0), (219.5, 184.0), (219.5, 185.0), (220.0, 185.5), (220.99999609374998, 185.5), (222.00000390625002, 184.5), (223.000001953125, 184.5), (223.999998046875, 183.5), (224.99999609374998, 183.5), (226.00000390625002, 182.5), (227.999998046875, 182.5), (228.99999609374998, 183.5), (231.000001953125, 183.5), (232.99999609374998, 181.5), (234.00000390625002, 181.5), (235.5, 180.0), (235.5, 178.0), (236.499998046875, 177.0), (236.499998046875, 176.0), (238.500001953125, 174.0), (238.500001953125, 173.0), (239.5, 172.0), (239.5, 168.0), (238.500001953125, 167.0), (238.500001953125, 165.0), (237.500005859375, 164.0), (237.500005859375, 163.0), (236.99999609374998, 162.5), (235.999998046875, 162.5), (235.000001953125, 161.5), (234.00000390625002, 161.5), (233.499994140625, 161.0), (233.499994140625, 160.0), (232.99999609374998, 159.5), (222.00000390625002, 159.5), (220.99999609374998, 158.5), (220.0, 158.5), (219.0, 157.5), (216.99999609374998, 157.5), (216.0, 156.5), (211.0, 156.5), (210.00000390625002, 155.5), (208.0, 155.5), (207.0, 154.5), (190.00000390625002, 154.5), (189.0, 155.5), (184.0, 155.5), (183.0, 156.5), (181.0, 156.5), (180.0, 157.5), (178.0, 157.5), (176.0, 159.5), (175.0, 159.5), (173.0, 161.5), (172.0, 161.5), (171.0, 162.5), (170.0, 162.5), (168.0, 164.5), (166.0, 164.5), (165.0, 165.5), (162.0, 165.5), (161.0, 166.5), (159.5, 165.0), (159.5, 164.0), (157.0, 161.5), (155.0, 161.5), (154.0, 160.5), (152.0, 160.5), (151.0, 161.5), (149.0, 161.5), (147.5, 163.0), (147.5, 166.0), (140.5, 173.0), (140.5, 175.0), (141.0, 175.5), (143.0, 175.5), (144.0, 174.5), (145.0, 174.5), (149.0, 170.5), (155.0, 170.5), (156.0, 169.5), (160.0, 169.5), (161.0, 170.5), (162.0, 169.5), (167.0, 169.5), (168.0, 168.5), (175.0, 168.5), (176.0, 167.5), (180.0, 167.5), (181.0, 166.5), (182.0, 166.5), (183.0, 167.5), (188.0, 167.5), (190.00000390625002, 169.5), (191.0, 169.5), (192.99999609374998, 167.5), (196.0, 167.5), (196.99999609374998, 168.5), (200.99999609374998, 168.5), (202.00000390625002, 169.5), (204.0, 169.5), (204.99999609374998, 170.5), (207.0, 170.5), (208.0, 171.5), (211.0, 171.5), (212.49999609374998, 173.0), (212.0, 173.5), (207.0, 173.5), (206.00000390625002, 174.5), (203.0, 174.5), (202.00000390625002, 175.5), (195.0, 175.5), (194.00000390625002, 176.5), (188.0, 176.5), (187.0, 175.5), (186.0, 175.5), (185.0, 174.5), (184.0, 174.5), (183.0, 175.5), (182.0, 175.5), (181.0, 176.5), (177.0, 176.5), (176.0, 175.5), (175.0, 175.5), (174.0, 174.5), (171.0, 174.5), (170.0, 175.5), (169.0, 174.5), (166.0, 174.5), (165.0, 173.5), (160.0, 173.5), (159.0, 174.5), (158.0, 174.5), (157.0, 175.5), (155.0, 175.5), (152.0, 178.5), (150.0, 178.5), (149.0, 177.5), (142.0, 177.5), (140.5, 179.0), (140.5, 180.0), (146.5, 186.0), (146.5, 187.0), (149.0, 189.5), (154.0, 189.5)]
pts = [(round(x), round(y)) for x, y in pts]
shape = (max([x for x, y in pts]) + 1, max([y for x, y in pts]) + 1)
img = Image.new("L", shape, 0)
draw = ImageDraw.Draw(img)
draw.polygon(pts, fill=255)
img.save('polygon.png')

x, y = zip(*pts)
plt.figure(figsize=shape)
plt.plot(x, y, marker='o', linestyle='-', color='blue', markersize=5, label='Points and Lines')

plt.scatter(x, y, color='red', zorder=5, label='Points')

plt.title("Visualization of Points with Connections")
plt.xlabel("X-axis")
plt.ylabel("Y-axis")
plt.grid(True)
plt.legend()
plt.savefig('points.png')

points

polygon

@SakuMyl
Copy link
Author

SakuMyl commented Jan 2, 2025

I was able to produce these issues by running
pip install git+https://github.com/python-pillow/Pillow
in my virtual environment after the recent merge to master.

@radarhere radarhere linked a pull request Jan 3, 2025 that will close this issue
Draft
@radarhere
Copy link
Member

I've created #8659.

As far as I can see, it resolves your issues. However, I imagine you have more examples apart from the three you've shared here, so I'm thinking of leaving it as draft until you give it your sign off. Would you be able to test the PR and let us know if it works?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
2 participants
@radarhere @SakuMyl