GrabCut.py

[molyswu]

Try to :
python yoloUtils/GrabCut.py --image ./images/person.jpg --yolo yolo-fish

File "yoloUtils/GrabCut.py", line 44, in
img, boxes, idxs = yolo.runYOLODetection(args)
ValueError: too many values to unpack (expected 3)

Thaks

[Cuda-Chen]

Hi molyswu,

Thanks for your response, I have fixed this issue.
The reason is that runYOLODetection only returns 1 argument, whereas runYOLOBoundingBoxes returns 3 arguments.

Anyway, I have pushed a new commit to fix the problem :)

[molyswu]

Hi， Cuda-Chen
Thank you for your help !
How to calculator pixel or size at grabcut area.

[Cuda-Chen]

You said to calculate the area of the rectangle, or bounding box?

[molyswu]

box

[Cuda-Chen]

In line 193, utils/yolo.py I write the following:
boxes.append([x, y, int(width), int(height)])
Also, in line 16, utils/GrabCut.py:
rect = (boxes[i][0], boxes[i][1], boxes[i][2], boxes[i][3])
Therefore, you can calculate the area of bounding box by multiply width with height (boxes[i][2] represents width, whereas boxes[i][3] represents height).

Maybe this helps :)

[molyswu]

Thank you

[molyswu]

bounding box of area ==fgbModel of area?

I want to get the object of area

[Cuda-Chen]

If you mean to get the foreground area after GrabCut,
maybe grabCut helps (GrabCut.py:20).

Shown in OpenCV doc, the grabCut function returns three objects: mask, bgdModel, and fgdModel.
I guess fgdModel is what you want.