search

gist-ailab / uoais

Codes of paper "Unseen Object Amodal Instance Segmentation via Hierarchical Occlusion Modeling", ICRA 2022
Other
130 stars 27 forks source link

Labelling the dataset #7

Closed jayes97 closed 2 years ago

jayes97 commented 2 years ago

Hi. I'm amazed with your work and I want to apply it on my own datasets. I want to know what tool you used to for labelling to get "visible_mask": RLE, # visible mask "visible_bbox": [x,y,width,height], # bounding box of visible mask "occluded_mask": RLE # occluded mask "occluded_rate": float # ratio between occluded mask and amodal mask all this things

SeungBack commented 2 years ago

Hello @jayes97 ! Synthetic dataset generation pipelines are following.

  1. Generate synthetic RGB-D images using BlenderProc.
  2. Generate the binary visible masks and amodal masks using boptoolkit. (We already have 3D object models and its 6d poses)
  3. Compute the visible_bbox, occluded_mask (A-V), and occluded_rate (V/A).

For OSD-Amodal, we labeled the amodal masks using supervisely.