Getting mask from Yolo bbox

[aimanmutasem]

Search before asking

• I have searched the Supervision issues and found no similar feature requests.

Question

Hello, and thank you for the great repo. : )
I'm following the example here for getting masks from coco bbox, I have tried to use the same technique to get masks from Yolo bbox format. One possible way is to convert Yolo bbox to COCO bbox as the following :

def yolo_to_corner_bbox(yolo_bbox, image_width, image_height):
    _, x_center, y_center, width, height = yolo_bbox
    x1 = int((x_center - width / 2) * image_width)
    y1 = int((y_center - height / 2) * image_height)
    x2 = int((x_center + width / 2) * image_width)
    y2 = int((y_center + height / 2) * image_height)
    return x1, y1, x2, y2

This technique does not work for me well compared to COCO bbox in the original example.

Additional

Is there any direct way to get a mask from Yolo bbox format?

[github-actions[bot]]

Hello there, thank you for opening an Issue ! 🙏🏻 The team was notified and they will get back to you asap.

[SkalskiP]

Hi, @aimanmutasem 👋🏻! Let me convert this question into a discussion and put it into the Q&A section.

[SkalskiP]

@aimanmutasem This notebook is already quite old, and since then, several new features have appeared in the supervision. You can now load your YOLO dataset this way:

>>> import roboflow
>>> from roboflow import Roboflow
>>> import supervision as sv

>>> roboflow.login()

>>> rf = Roboflow()

>>> project = rf.workspace(WORKSPACE_ID).project(PROJECT_ID)
>>> dataset = project.version(PROJECT_VERSION).download("yolov5")

>>> ds = sv.DetectionDataset.from_yolo(
...     images_directory_path=f"{dataset.location}/train/images",
...     annotations_directory_path=f"{dataset.location}/train/labels",
...     data_yaml_path=f"{dataset.location}/data.yaml"
... )

>>> ds.classes
['dog', 'person']

[aimanmutasem]

Thank you @SkalskiP, this works for me ... : )

[SkalskiP]

Awesome! That's what I wanted to hear!