Cuvis data Labels2TV does not accept segmentation in RLE format

[muehlbauer-cubert]

COCO expects RLE format as segmentation if there are multiple objects of one label in an image.

The data converter to TV does not accept this format.

post_it_segmentation_annotations.json

with open('jsonfile') as f:
    d = json.load(f)
    for ann_id, ann in enumerate(d['annotations']):
        img = d['images'][ann_id]
        size = img["height"], img["width"]
        test = Labels2TV.convert_COCO2TV(ann, size)

[manke-cubert]

Fixed by https://github.com/cubert-hyperspectral/cuvis.ai/pull/40