I was working with the code process_kitti.py by coxlab from GitHub in an Anaconda environment. Some of the function was deprecated in Python 3.6. Therefore I have changed the following line:
im = imresize(im, (desired_sz[0], int(np.round(target_ds * im.shape[1]))))
into
from PIL import Image
im = np.array(Image.fromarray(im).resize((desired_sz[0], int(np.round(target_ds * im.shape[1])))))
Otherwise everything else remained the same.
Interestingly, while I ran this code, the following error:
Creating train data: 41396 images
Traceback (most recent call last):
File "process_kitti.py", line 104, in <module>
process_data()
File "process_kitti.py", line 84, in process_data
X[i] = process_im(im, desired_im_sz)
ValueError: could not broadcast input array from shape (424,16,3) into shape (128,160,3)
I am a bit confused with the cause of such error. Thanks a lot with the help.