Creating SBD dataset from polygon information

[ogail]

Hi, I was able to successfully train and test on PASCAL VOC dataset from scratch and score accuracy of Mean [email protected] = 65.5%. Next, I want to conduct FCIS training on my own dataset, however the object boundaries ground truth information I have is in polygon points + class format like this: {class: "pen", "polygon": [(x1,y1), (x2,y2),(x3,y3),(x4,y4)]}. I'd appreciate if anyone is aware of an existing code or an easy way to convert this into SBD format. Note that closest thing I found was converting SpaceNet getJson labels to Pascal VOC labels (however, this requires converting what I have to getJson labels, which is still not as straightforward).

Any help/tips on a way to easily convert polygons/class to SBD would be appreciate.

[chrieke]

Just convert to COCO json format, that is much easier and FCIS also works with COCO. Conversion to VOC format is a lot more difficult and there is no library etc for it afaik.

[zhanghui-hunan]

@chrieke Thanks for your great work！i want to convert COCO json format, too. But i follow COCO_2014 dataset which the annotation file need instances_valminusminval2014.json and instances_minval2014.json, i cannot find any way to get my own .json file I mentioned. If you have a good way to share it with me? or your own dataset file structure, grateful!

[chrieke]

@liangzhicong456 Hi, I finally uploaded my preprocessing script for the image cutting and cocojson conversion. Currently adding the rest of the project code and will convert the preprocessing code to a nicer notebook with more annotations etc. Hope this helps you out in the meantime.