You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
It's a very small issue but I do think Fury should either give a warning or error out like PIL did when dealing with something like this. Could've saved me half an hour. I'll try to go through the source and work a PR, or someone else can also do it before me.
Thanks!
The text was updated successfully, but these errors were encountered:
Hi, this issue could be solved by adding dtype checks to all the texture functions in the actor.py file that only accept rgb image data of type np.uint8. Specifically adding the following exception raising to the texture(), texture_2d(), texture_on_sphere() and texture_update() functions:
ifrgb.dtype!=np.uint8:
raiseTypeError(f'Can only handle images of type: np.uint8. DType of the given rgb image is {rgb.dtype}')
Hi, I want to be part of this organization Google Summer of Code GSOC 2023! Can you tell me some details about what I should do to be accepted, I would really like to take part in this program and be able to learn as many things as possible. I made a code for this problem, you can take a look.
import numpy as np
from fury import window, actor, io
def check_image_data(image):
if not isinstance(image, np.ndarray):
raise TypeError("Image data must be a numpy array.")
if image.dtype != np.uint8:
raise TypeError("Image data must be of type np.uint8.")
def flip_image_horizontally(image):
h, w = image.shape[:2]
new_image = np.zeros(shape=image.shape, dtype=np.uint8)
for i in range(h):
for j in range(w):
new_image[i, j] = image[i, w - j - 1]
return new_image
This is a small issue, but it did cost me some time. I manually tried to flip an image horizontally by creating a separate numpy array.
Ok! So here' what I got:
This baffled me for quite a while. After a few minutes of debugging, I tried:
And PIL caught the problem:
By default, at least on my system,
np.array
chooses to use 64-bit floating point numbers which seems to be an issue. Thus, changing this line,to
Fixes the issue and gives the desired result.
It's a very small issue but I do think Fury should either give a warning or error out like PIL did when dealing with something like this. Could've saved me half an hour. I'll try to go through the source and work a PR, or someone else can also do it before me.
Thanks!
The text was updated successfully, but these errors were encountered: