Closed li-zheng-hao closed 4 years ago
Which one is rotated in the images you present? The gray background image (the real image) or red annotation?
I update labelme version from 3.16.0 to 3.18.0. And here are my images:
origin image
generate JPEGImages folder image
generate SegmentationClassPNG folder image
generate SegmentationClassVisualization folder image
the code is copy from labelme2voc.py
@wkentaro
and all images' shape is 4000×3000 ( width × height )
Can you possibly upload related files (original image, annotated json file, exported files) as zip? You can upload zip on GitHub comments.
The JSON file is missing?
sorry ,I forgot it ...
I reproduced. https://github.com/wkentaro/labelme/pull/543 will fix this.
thank you very much
Problem
i used the code in
labelme/examples/semantic_segmentation/labelme2voc.py
and the output result is not correct: I'm sure there is no problem with my markup,However, the output labeling result is incorrect.labels_file
inlabelme2voc.py
, i just create the file and input my labels manually,like:my env: