Is there a quick/batch way to convert segmentation bounds to detection bounds in a detection project?
do you mean converting segmentation into Bounding Boxes ?
if yes , then…
a Three dots at the Upper Right Direction click on it and you will find methods to convert masks into polygons and other solutions
this is intended if you already uploaded the dataset into roboflow
if not just implement it yourself then upload it
The annotation editor has a “Convert…” submenu (via the image ... overflow menu) with these options:
| Action | Scope | Location |
|---|---|---|
| Polygon → Box (individual) | Single annotation | Right-click context menu → convertToBox() |
| Mask → Box (individual) | Single annotation | Right-click context menu → convertToBox() |
| All Polygons → Boxes | All annotations on current image | ... menu → Convert → convertAllPolygonsToBoxes() |
| All Polygons → Merged Boxes (per class) | All annotations on current image | ... menu → Convert → convertAllPolygonsToBoxesMerged() |
| All Masks → Boxes | All annotations on current image | ... menu → Convert → convertAllMasksToBoxes() |
(there is no batch/bulk way to do this)
Is that what you’re looking for?
Ah thank you! Yeah the Polygon → Box.
But is there a bulk conversion for all my images?
you then should do Automation , use playwright
you can check my linkedin I did a relative task using Automation with Python and Playwright
Linkedin Post : Automating Image Annotation Workflows with Roboflow | Ahmed ElBamby posted on the topic | LinkedIn
Github Repo : LinkedIn
This topic was automatically closed 21 days after the last reply. New replies are no longer allowed.