Semantic segmentation of bicycle tires

[honey-zhao]

Hello, deeplabv3+ is a fully supervised algorithm and I would like to know how the official segmentation map of the bike is obtained? In other words, how is the ground truth of this internal skeletonization map labeled? Thank you！

[Tushar Patil]

Hi @honey-zhao,

These segmentation masks are obtained through the tiresome annotation method. We can use plenty of open-source tools available online like labelImg, vggannotator, etc to annotate. You can find similar segmentation masks along with original images provided by different open-source datasets like Pascal VOC, Cityscapes, MS COCO, etc.

Hope this answers your query!!!:)

[honey-zhao]

Thank you very much! I will try it.