Background of image turning black when using keras.preprocessing.image load_img

Question

my code:

from keras.preprocessing.image import load_img, img_to_array
from keras.preprocessing.image import smart_resize
import numpy as np

#load pretrained model
os.chdir('path/to/dir')
# model = load_model(model_path)
input_imagesX = natsorted(os.listdir())
for idx, v in enumerate(input_imagesX):
    
    img = load_img(v,color_mode='rgb')
    img = img_to_array(img)
    img = tensorflow.image.resize(img, (128,128))
    # img = np.expand_dims(img, axis=0)
    plt.imshow(img)
    plt.grid(None)
    plt.show()

Here, the transparent background of the png turns black and the fill gets distorted. Have tried a couple of hacks, none seems to work! Can someone please help?

original image

image i get as the output

Answer 1

Some images might also have an alpha channel. Please find below how to load the alpha channel

img = load_img("/content/transparent.png",color_mode='rgba')
img = img_to_array(img)
img = tensorflow.image.resize(img, (128,128))
img=tensorflow.cast(img,tensorflow.int32) 
plt.imshow(img)
plt.grid(None)
plt.show()

Output:

Let us know if the issue still persists. Thanks!